1, J GRAPHIC DESIGN PRINCIPLES POSTER 48°74'14"N 19°15'13"W Definition of graphic design Graphic design principles Typography Process of creating a poster Case study Graphic design DEFINITON Graphic design Definition * LadislavSUTNAR Large printed format consisting of text, image, photography POSTER On public places To inform Format 841 mm 1000 mm 917 mm t E E a* 00 A8 - A6 A7 A4 A2 A5 A3 HU A1 E E 5- IB3 B7 B6 B4 B2 B5 B3 D B1 E E a* C8 C7 C6 C4 C2 C5 C3 r CI Format Ao, Ai Science posters and technichal drawings Ai, A2 Working boards (at meetings) A2, A3 Diagrams, drawings and big boards A4 Magazines, letter, leaflets,... A5 Notebooks, calendars A6 Postcards B5, A5, B6, A6 Books C4, C5, C6 Letter enuelopes B4, A3 Newspaper B8, A8 Playing cards Science posters Case studies Real-time Analysis of NetFlow Data for Generating Network Traffic Statistics using Apache Spark in SS3 • "■ o Verification of Programs with Inputs DIVINE Symbolic States tÜ EACirc Using genetics to improue encryption Vlartin Ukrop, Petr Svetida. Marek i,-:. Vaclav MatyäS et alii Problem statement EACirc workflow 1. Forming d population t » I EMit a fn 5. Mutation S cranDuer 3 C WTSf X I 4Suruiuilo-»E5^H iput variables and a »OH3D-»<0>->8 le current DIVINE, this fffl-H M-WlH^|>.i( configurations. In sym- Verification algorithm Correct Error the transformation can also be used for different representation of symbolic data quite easily. The transformation is handled by LART - LLVM Abstraction & Refinement Tool. Furthermore, DIVINE: verification algorithm has to be ofTymbolicPstates. For both of these tasks, DIVINE has to extract SMT formulae from the program state and use SMT solver. OJE) , Dependence graph of mtrolflow ofaprog ESferaDiSe MEiiCirc Using genetics ti improue encryption Martin Ukrop, Petr Švend , Marek Sýs, Václav Matyáš et alii Problem statement Randomness testing M The ciphertext produced by |jir!P ▼ flk ft J* £ == t jk im\ • ♦ t IM«»' Iteratiue design EACirc workflow 1. Forming a population cipher data from created randomly. t 2.Testuectorgeneration rt ^ |^| I *t Testingdataforlearnmgis | ^ ^ r 3. Fitness assessment 1 W K W "ST # -1 ♦ + + * * 5. Mutation & crossouer it. Suruiual of the fittest Comparison to existing tools EACirc us statistical testing rt The standard wa^to assess randomness ,s to ■■ analyzed 77 different functions {eStream, Sr-1 -round Fubuki we confidently surpass CRwCS EACirc ✓ Further information • ^ below or ask directly at the lab (CRoCS @> Fl MU). DÍehard(er) ■■>. ^ vol.-ase isecrvpt zd13J. springer Berlin Heidelberg. 201-a. ... x ;r.r;r.™rrr-rr,*"'—■* Colour PURPLE secondary Meaning of colour is always completely accepted subjectiuely. The perception depends on two imperfect human organs - eyes and brain, depends on light waues. Colour has a powerful function in graphic design. Colour defines hue, saturation and lighteness. V. 70S [power Idanger Uassi°n % 1 ° 5 3? #0* 5^ mature i renew ali tranquility! HOPEt 3^ .1 > > m p i Jflfl secondary ORANGE Colour Warmer colours appear heauier than cooler colours Contrast Relation between two different elements colour contrast contrast between elements High-Intensity colours appear heauier than low-intensity ones positiue and neagtiue space geometric and biomorph shapes softness and sharpness stillness and mouement big and small Contrast Elements and background The higher the ualue-contrast, the heavier the weight of the object White space in Spain ousing • government quarters • rental guaranty • on-the-cconomy «N ANY perrflarwnt change of ma-lion, htiustni:, or rather iU avaita-tiilily. ■■■'■■.•me* I in? most Important question y ked by llii; service fasjüy. The conwnon problemi you ibmi In the United Suite* mi- wjmewlmi ninj:ntrir.'l en an overseas movement beeniwo of diJFpnriüK in taiwiiasr, law htv* niMnrn. Acmfilly. At leant FcT Sjwtn so our expo-rknecs hnw proved, th(»e problern.« aren't very Wb lit oil it you're- toM br>ftiri4iatn3 what 1u expert. First of all. st the time Mm booklet ■was published, jiuttotmtk concurrent trawl i.F dPncndcflU la Spain *vn« authorised only far cv^mmJ» arid general officers. All other military piriQUIHl must apply to live appropriate ov«n>aea SpCCiflC III •* i li. 1. -ii. on liow to s! j lKi« can be obtained from your personnel of* flour, tVlien llir isi.'^s timum rider Brants approval lor concurrent travel, he will Lell you whether "wenaufnl quarter* me or aw not available. And. of nuiE, this Will determine many cif yonr ■-! i'".' ■!»■■■: ' net torn Government ojiartcni con*M of cn-bAM* and renlnl rriinninty hnusire., tome Whit ftnYJIar (O 41 ID sfrrtdlcd Whrrry hi«i"in..: in (he United State«. On-base huuiins 0,1 the Air Koree Ijnsrm ■- very Mmlt«Jr tannine from 2C-10 ttrtitt and tu mirletad ta key iwwnrwh The Rota Nav.nl Bnso. where Immisiiim in the levni eommuilltki Is rxlmrx-Iy llnuEAil, Itwrp m» 19U cin-he.se unit*. There, is no rental gwranty hetiwljiji In the Kd(B area. Korly-alx unlut are under const ruci kin at the Cstrtnuefiii Ctanl Facility. There ore none til I lie £1 íerrol Naval Facility. There are 20 Units each, si the, nir. craft control j will wamlnjc *He* at Villa, lobar nv-ai uxl Constant! aa iW-S). Twenty will* are unrbjr ccMwtniclJw) at Rjti tW-l) and itaúdorru íw-&). th™. inn N under desljpi for icme of the other sites. HK4TAX OrjAHAHTl HOUSING In the Madrid area (Uifc* Includes the Joint U. S MHitaiy CI roup. Spain: KAAG; NAVACTS. Spoil.; llntdquar-ten Sixteenth Air Korce; I Lm-I .um r-•era. Both Air WvEsipn; Torrctfri Air Eiiae and several smaller unlla I hero arc SK hoiuunjc units, cnllrd Itaynl Oak*. (a«ited five- mtlc* liorlh of Muilrid and ttppn>Ni mately So mllofl from To-rrcjon Air Raoc. In ZnrarKHa. there? arc 232 units; In SevbLU iWurtiii nn:L Sun Pablo nir basa) . there are 4&& units sboul one nula iroan ihe dly. All units ore spacioua, allhwiKh me bodroomi aea som«tvliiit umíiller than Atiterlciin standarde since emphasli hat been placed tin tb* llvlng-dilnin^ area., A typical |w>(HKMlf(Kint unit hn-s a larjpe Irrrprr, flvinn; room. ťlinLilíI room, matter he-:ironm, » Mnnllcr bedroom, btlli, hitchnv, utility room, Btamse tujm, mild n nvild1* room br<1 baih. Tlie la_ťRír unit* arc baiiťally the snmc. If yiiu ure notifíecl lltat ynti nr« to brettpy flůverfiřntni quaríem—cither on-base or rental ppmrariiy- -you will be allowed to riiin only SffK pountb or liouiielioli Ft«Kis. pirn your hold Imir-ES1ÍC .Uiu fcuid hiKKHi!'*!- All Ktwerrinreat niiiiiriers nre oiteijuatcly and eontforta-lity furnlsbeii, im ln.l.n;: Move, refrljte-rator, AUlomntl? wiixher-drycr mmbirta-tinui. vacuum rlcaix-fii, rug* and drape-rt«. AJwi Included tire lampi, waiio-papei' 3n»skrtt*. porch furniture. Ironing Hard, etc. Prrsniitirt bolm; nssipicd to Rota Nn« val Dose ihcuhl note lluil ihi-c unlta do nm ineludíř wathliqí nuehlnes or rlothex dm j^. Normally, you will need hTinp; only dishe-.. KUwrwarc, poti and panu, IUh-u 4ltieJtiitinK pilli™< and blanket*), and personal itenu. You will probfltily wanl In hilni[ yiiiir unall npfilLaiicex—iron, mixer, i... — . i : ii;., all tienu of this type work wM Im Sixihi and will *»V« ytisi ax imirh «-iult iij tfiev {|o In the United Slatex. You ahonld indnde in your 2.000 pcmiMl weiehl limit all snceial Lt-nw taw beibks find stnali thltd-ren since no Henu or thl» iKitore arc furnished. Thl» wimiM Iwhirte crUW, youth lwd» (It you us-, them), vnpgr-i£ěri, botUe si«rliix«fs. ete. As a ml-W Of fact. If you ore ImvcJJnE With « bottle-baby, vm suggest you include in your Ii-ipkI basKaiee tthnl is, brine Jt with yon) r hn|t|i> sterilizer—the type yi m can use on Ihp top 1,1 " stove. By the way. nlusllc bo!tie. me much raore practical, lihrluile pxtm nipplrt, CansUkr your upcm equipnwtn. ctuliU (Vn's boyi. etc.. In the 2.0GO potintis. Hold hjijjjy«KC, wWch wJII arrive much MMorr than yaur fumilure, should Include till KW ilL-irn you Will iiniiiLYljiitely need. Wc tiii.-i.il :h\:. meant llw baby crib, ii*tn» toys, a trwl kit ChammeT, saw, pliers scrrXVtlrJWJ-s, tic.*, dishr*. pt»t« and pan* aim) other ceoktaa; para-pi.L-;ii.-.l::i. silverware, Ihncns, bkiiikebi -»-enoU(-h In set tip lomporary liouto-kcejrin^ for nl-mt dx In rifthl Weeks, In our luiint bamiaKe.uthrr Ihun eloth- 1\>K which Is iUffcuAflrd ctMwhere. we Inrlnned extm tooth pn*te, rai-or bJn-ill--.. At leatt one ley nee- ehlld. iM-.fi *i\\v*r flhWill perfOnal Items wliltli yoru fjiiir.fH ennvenlenlly buy white enreule. OH ECO-XOalY HOUSING Uvlfui on the econonty. accorulnii! \o the mnny Americans who do to. pro-vi v-. :: in.My :i[iri :n.i,.■■[■.-,ijyir.tr,i,l to the American way of life. True, thc- dlfferenren nn> Jawnctkmej trtl^tf-ating, butt they are usually minor, and don't detract from the opportunity to learn 11» ton&UR£e find tlítíontt of Sjaiň. Whether you cvcni uaiiy choose a house (of which there are very fewl or an White space, unprinted space without any element which surrounds other elements to make design more legibile and lighter. White space Ladislau Sutnar A typical iwo-Uslrmrn wilt hrm a lnrxe lerrarc, rtvintr room, cUntnji room, matter bedroom, a smaller bcriratwri. btlh, hitfJieti. utility room, storage rowi. and maid's, rocm nnd Isnh. The bLfjar nils are basic-Ally the snmc. ' you are, tiub'fird Hut you ar*> to eowmmem quarter*—either mi r rental xiHJtraniy you will be allotrcmlo «hjlp only 2.000 pounds or liuumtfunV poods, plus your held bfljf-I'-ir.i. ,i.ii.l^ki;i',l hiKKHirn. Alt covennrent Hiuiirtcrs i« nitcijuatejy and eemforta-hly (taiil**!^, inrtiirlins stove, rvfrdcrf-ish^r-rfryer wmtana-ft-aixni, rue* and drape ' ere Lamps, wtilc-i furniture, iroMlms ird, etc. ■ Personnel beMi; iisaiiytetl to Rota iVa-vnl Base shciiij oot<> Haul tlie-'i- wills wtuhljq; machines or ON ANY poiTttanerit Cfianec Of iU lion, huusüfcK. or rather Us availa litily. become* Hi,- m.i.i Important question asked by llie serving- family. The common problem* you iumi In the United Slain arc somewhat magnified □n nn wm«oj movement bmu» of difrrrennw Irt hnxuaxr, law nnri niatoni. Actually. At leant for Spain to our experience* have proved, throe problem* nrcn'1 vr-ry V-:--: nt all if you're told 1..''.i.-i/:i>:í.I Wh.il !:■■ .-.!■. Find of nil. at the time this booklet was published. ■• ■■ •> ■ concurrent travel o\ dependent* lu Spain Was authorized only for eölonwl» ojkI general ofíkcri. All other military pei-sutmcl muit apply to üwr appropriate ov«n*aa cnmi"»rd*r (or concurrent Irivl. Specific Ins tract kins on 3low to do this can he obtained from your personnel of* fleer, When tlie tvcrsuras euuLria rider (rants approval lor concurrent travel, he will tell you whether KOVcrnmeiiL quarters iirc or nro not riva liable, And. of court?, (his. will determine many úT your -ii'im quei ' actions, Covtrnrnent nnnrtcni consist of cn-bťi*p ani) rental rcitnmnry housing, somewhat jiÜilSlar (Ol (He xu-tfilled Wherry Sivi-in : in the United States, On-han* hMUinic ai the Air l1,,-. .■ |R very BrviltM, rniirjlns From 2C-10 úniu and Is. PT«tr1eH>iJ In Uty nerton rwl. The Rata Htm Enw. whir© Imuiltw 10 ^ihe local eofnrouiiltks use iinitR. There. hn^^HrtS^iuiininiy howslAji In the Kdta area. Korty-ittx unlls are under cor.si ruciitin ni the Oart-nRCfin Naval Kir!"Ely- TIiitl- none al the £1 Pcrrol N.,vnl nieUiiy. Itwre are Í!Ď UilHi each qI Ihn ntr. craft wnlral innl wiimiiit Mlea at Villa-lobns tW-3> ojkI Comdanlina . Tmily uiillx Mrr unrl^r cgrwtnictlwi at R^as iW-li »rui BiHŇJot-iii i! uppi-osiritatcly 20 miles from To-rrcjůn Air Rate. In ZarnrřtHa. there arc 222 units; In Sevilla iNk-r-if; and Sun I>afato ojr Ssi-sei). there arc 4S4 unita uboul one ntilo from the dly. All units arc spaclog*, althwieti lbe bedroomi arc aomeivhat smaller than Anti.TÍr;in *tandnrd> aince emphails has ban placed ún the lis'lns-dlnin; areas. Will m.'c, tir.n ; only P*ti aixl ponK, liinn« .«nn hlanhcts), and 'ou wi]] prnbflljly wnnl small appliance*—iron, - m i i-i 11> all Items «f well tn Spain and vvUI rnwh wmrlt A3 Ihey stcrfliacr— the type ymi can use cm thi> top i>l a stave. By the vray. |i|iis!ic bottlet tiie much more pnxetlcuL I i»:-3ipxtm ni(iple». ConsUkr your sports equipment, chihu ren'B toyi. ete.. in the 2.000 pounds. Hold luu^ae,"'. which wJll arrive much Kth-Jcwr than ycuir furniture, ^Hostld Include lltlMO i'cirl-. you Will tttimCiliiiU'hr need. We found this meant IIm baby crib, ain» toys, a loal kit (hammer, saw, filler^ screwdrivers, etc.*. dishrj*, pots and pans and other ecokinK para-phcrnaiLfl, silverware, linens,, blanhebi — enough to set ap temporary Ixhdc-liceptnM for alMKit kx In eijrhl week*. In our hnitd hn,ir(5iRe.-rjther Ihnn cloth-Iob which ts dJnctiMed elacwhere, we tneluded cwtrn tooth paste, raior Wb-des, at least one ley per child, nnr! oilier nnwill rjersonal Items uldtli yi>n cannot conveniently buy while enroule. C1N ECON'OMV IlOUSlNtl i ivn-l-r on the economy, acrxirdfng id the mnny Amerienn* who do so. provides a lively And iitirn-itins contrast to the American way of lite True, the dlffcrmren are sometbnet frustrating', hut they arc usually ininw, him) don't detract from the opportunity to learn the tenguase and eusttmis of Spain. Whether you eventually choose a rnuse tot which there are very few) or an 50 % 1 25% 25% White space Lorem ipsum dolor sit arftet, consec-tetur adipiscing elit. AeneSm neque quam, interdum non aliquot non, sagittis vel sapien. In risus puriX rut-rum aliquam massa id, elemem vestibulum dolor. Curabitur id doŤ tellus. Donee interdum mi purus, ei maximus mi efficitur eu. Vestibulum porttitor est in venenatis eges-tas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consec-tetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rut-rum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donee interdum mi purus, eu maximus mi efficitur eu. Vestibulum porttitor est in venenatis eges-tas. Morbi placerat maximus suscipit. Proin interdum diam massa. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum al- iquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donee Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, ele-Lentum vestibulum dolor. Curabitur ictalolor tellus. Donee interdum mi puots, eu maximus mi efficitur eu. Vestibulum porttitor est in venenatis egesteis. Morbi placerat maximus suscipit. Kfoin Lorem ipsuSn dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdu\n non aliquet non, sagittis vel sapienVtn risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus.Proin interdurn^diam massa. Consectetur adipiscing \lit. Aenean neque quam, interdum rWi aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa kL elementům vestibulum dolor. Curabitur id dolor tellus. Donee interdum Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efficitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efficitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin interdum diam massa. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam mas- White space between elements White space Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi plac-erat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi plac-erat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum iiquam massa id, elementům bulum dolor. Curabitur id dolor í^llus. Donec interdum mi purus^u maximus mi efncitur eu. Ve^bulum porttitor est in venenatis °gestas. Morbi placerat maximhft suscipit. Proin interdum diam^riassa. Lorem ipsum dolor sit ín^et, consectetur adipiscing elit. Ae nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus.Proin interdum diam massa. Consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, el-mturn vestibulum dolor. CuS^bitur id dolor tellus. Donec interoHm mi purus, eu maxi-grdum mi purus, eu Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin interdum diam massa. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus.Proin interdum diam massa. Consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi interdum mi purus, eu White space Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi plac-erat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi plac-erat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ae-nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum iiquam massa id, elementům bulum dolor. Curabitur id dolor í^llus. Donec interdum mi purus^u maximus mi efncitur eu. Ve^bulum porttitor est in venenatis °gestas. Morbi placerat maximhft suscipit. Proin interdum diam^riassa. Lorem ipsum dolor sit ín^et, consectetur adipiscing elit. Ae nean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus.Proin interdum diam massa. Consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, el-mturn vestibulum dolor. CuS^bitur id dolor tellus. Donec interoHm mi purus, eu maxi-grdum mi purus, eu Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maximus suscipit. Proin interdum diam massa. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aenean neque quam, interdum non aliquet non, sagittis vel sapien. In risus purus, rutrum aliquam massa id, elementům vestibulum dolor. Curabitur id dolor tellus. Donec interdum mi purus, eu maximus mi efncitur eu. Vestibulum porttitor est in venenatis egestas. Morbi placerat maxi- Case study White space Jan Tschichold Die neue typographic 6839205^020173083483 67542003291668 98367154 Case study White space 1 FACULTY OF INFORMATICS Muuyk Unmnily Verification of Programs with Inputs Henrich Lauko, Vladimir Still, Petr Ročkai, Jan Mrázek and Jiří Barnat DIVINE n DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). A big downside of the current version of DIVINE is that for programs with inputs, this input has to be simulated by nondeterrninistic choice which is very inefficient. Therefore we present an approach for symbolic representation of inputs in DIVINE. Symbolic States Consider a simple program with 32 bit input variable x and a branch on the value of this variable. In the current DIVINE, this program gives rise to 2iz possible memory configurations. In symbolic version, possible values of variables x and b are represented symbolically using bitvector formulae, therefore, there are only two possible configurations at the end of the program. Explicit Approach >QE23-»«^5>-»| »Q2£J-»4^>-»| Proposed Approach SMT Solver Verification algorithm Interpreter 1 Correct ► Error To take advantage of symbolic representation of states, we transform the LLVM bit- in detail later. Apart from that, the verification al codein such a way that it represents variables which can contain values dependent on symbolic states with the help of an SMT solver, inputs symbolically. This transformation is performed by LART and is resented irithm is modified to handle Our approach aims for minimizing changes to the LLVM interpreter that is used to execute instructions in DIVINE The reason is that the interpreter is complex and performance tuned and therefore it is not desirable to make it even more complex by adding symbolic data manipulation into it. Instead, symbolic data are to be handled by the program itself. To encode symbolic manipulations into the program we transform the LLVM bitcode produced by the compiler and create of symbolic states. For both of these tasks, DIVINE has to extract SMT formul; symbolic LLVM from it. This not only minimizes changes to the interpreter, but from the program state and use SMT solver. the transformation can also be used for different representation of symbolic data quite easily. The transformation is handled by LART - LLVM Abstraction & Refinement Tool. Furthermore, DIVINE'S verification algorithm has to be modified. It has to check if symbolic states are valid (nonempty), that is if they can represent at least one concrete state. It also has to handle cornpari: Details of Program Transformation n- I LLVM j LLVM bitcode is generated from C++ source code. Dependent instructio tuted with symbolic calls, path c< ipulations are added. LART takes the LLVM bitcode of the program and libraries produced by the compiler and transforms it into a bitcode which manipulates data symbolically. In this modified program, any variable which can depend on an input value is represented symbolically using bitvector formulae. Bitvector formulae describe integers of fixed bit width with overflow and bitwise operations, and therefore nd therefore bohc variables rig ParaDiSe Parallel fc Distributed I_LIU Systems Laboratory are well suited for exact representation of computer integers. All the manipulations with such variables have to be transformed to their symbolic versions which modify the formulae accordingly. Furthermore, any branch which depends on an input value has to put constraints on the possible values of symbolic variables (this constraint is given in the form of a path condition formula). white space Case study White space without white space Real-time Analysis of NetFlow Data for Generating ^jbrs/^ Network Traffic Statistics using Apache Spark Milan Čermák, Tomáš Jirsík, Martin Laštovička Institue of Computer Science, Masaryk University Botanická 68a, 602 00, Brno, Czech Republic ^NA^**1 E-mail: {cermak, jirsik, lastovicka}@ics.muni.cz Abstract — We present a framework for the realtime generation of network traffic statistics on Apache Spark Streaming, a modem distributed stream processing system. Our previous results [1] showed that stream processing systems provide enough throughput to process a large volume of NetFlow data and hence they are suitable for network traffic monitoring. This demo describes the Integration of Apache Spark Streaming into a current network monitoring architecture. We prove that it is possible to implement the same basic methods for NetFlow data analysis in the stream processing framework as in the traditional ones. Moreover, our stream processing Implementation discovers new Information which Is not available when using traditional network monitoring approaches. Systems Performance Benchmark — Four Nodes (32 vCPUs) ■ Samza and Spark have a high-enough flow throughput and can be used for the analysis of data from multiple networks at the same time. ■ Apache Spark system has been chosen as It offers an easy management and a high versatility in terms of the running environment and proprietary processing methods (e.g., sliding window). min in íl Framework Architecture The demonstration duster consists of 7 virtual machines, one is dedicated to IPFIXcol, 5 to Spark and one to the Kibana and Web server. The following configuration is the same for all machines: ■ (4 vCPUs) Intel(R) Xeon(R) CPU E5-26S0 0 @ 2.70GHz, ■ 8GB 1600MMHZ DIMM DRAM EDO, ■ 85GB SCSI Disk with 53c1030 PCI-X Fuslon-MPT Dual Ultra320 SCSI, ■ 10Gbit/s network connection,! G bit's virtual NICs. IPFIXcol is a flexible IPFIX flow data collector designed to be easily extensible by pluglns. In our demonstration, we use only part of Its wide functionality - data acquisition from multiple network probes and ttieirtransformation intoaJSON data stream. :istics Volatility of Network Traffic Data I Stream processing provides more m par 5 second Interval volatility in results compared to traditional flow data processing approaches. I Allows to observe short but , strong bursts of the network traffic that were lost due to the aggregation used in traditional batch approaches. Real-time TOP K Statistics Provides a real-time computation of Top K statistics to enhance network situational overview. n per 5 minute Interval • I #f0i í , SecurityCloud Acknowledgements [1] M. Čermák, D. Tovarňák, M. Laštovička, and P. Čeleda. A Performance Benchmark for NetFlow Data Analysis on Distributed Stream Processing Systems. In Proceedings of NOMS, 2016. This research was supported by the Technology Agency of the Czech Republic under No. TA04010062 Technology for processing and analysis of network data In big data concept. All elements should co lie rate together typography graphic style of elements grid Case study Graphic design style san serif typography left and right alignment MASARYK jj UNIVERSITY li EACirc Using genetics to improue encryption Martin Ukrop, Petr Švenda, Marek Sýs, Václav Matyáš et alii Problem statement Randomness testing The ciphertext produced by encryption should be completely indistinguishable from random data. But how to compare? EACirc is a framework for designing a distinguisher - a simple program that decides whether generated ciphertext looks random enough. EACirc workflow 1. Forming a population A set of currently considered partial solutions (gate circuits distinguishing cipher data from random data]. The initial population is created randomly. Z. Test uector generation ^ ing data for learn sampled from both sources. That is, non-random data from the inspected cipher and random data from a truly random source. 3. Fitness assessment Each circuit from the population is evaluated on all test vectors from the current set. Based on the outputs, it is assigned a fitness value from the interval [0,1]. i—i—i—i—i • ♦ ♦ ♦ i 0.792 0.503 0.357 0.S45 0228 He rati ue design The designed distinguisher is in the form of a gate circuit (layers of simple interconnected functions]. It processes binary data and outputs a randomness verdict It is improved iteratively, using ideas from evolutionary algorithms (see the next section for details]. 5. Mutation & crossouer To form new individuals, we use makes small random changes in nodes and connectors. Crossover creates an offspring by combining different parts from two circuits taken from the population. The new population now enters the evolution cycle again, gradually improving its fitness. k. Suruiual ofthefittest Unfit individuals are discarded, better ones survive to the next generation. The higher the fitness, the bigger is the chance of survival. The evolution works as a heuristics looking for better individals - gate circuits distinguishing random and non-random data with higher probability than random guessing. Style of typography Comparison to existing tools EACirc us statistical testing C\ (EACirc ^ ✓ Further information The standard way to asses jse batteries of statistical te= Dieharder or TestUOl. We n EACirc and c randomness is to s such as wsrsre, n them along with iinpare the results. To have a fine-grained comparison, we have analyzed 77 different functions (eStreom, SHA-3 and CAESAR candidates]. For 2-round Hermes and 1-round Fubukiwe confidently surpass NISTSTS. EACirc L.NIST-».* Diehard(er) f !-► ■■■ ■+■ x Interested in EACirc? See the papers referenced below or ask directly at the lab (CRoCS @ Fl MU). ainguishers foreStrewn nutí: >I.456[SECRYPT 2013). Sprir CRaCSě et GAP 2 02/11/0422. Case study Unity same styles of arrows same colour of background shapes same weights of strokes MASARYK jj UNIVERSITY li EACirc Using genetics to improue encryption Martin Ukrap, Petr Švenda, Marek Sýs, Václav Matyáš et alii Problem statement Randomness testing The ciphertext produced by encryption should be completely indistinguishable from random data. But how to compare? EACirc is a framework for designing a distinguisher - a simple program that decides whether generated ciphertext looks random enough. EACirc workflow 1. Forming a population A set of currently considered partial solutions (gate circuits distinguishing cipher data from random data]. The initial population is created randomly. Z. Test uector generation ^ ing data for learn sampled from both sources. That is, non-random data from the inspected cipher and random data from a truly random source. 3. Fitness assessment Each circuit from the population is evaluated on all test vectors from the current set. Based on the outputs, it is assigned a fitness value from the interval [0,1]. i—i—i—i—i • ♦ ♦ ♦ i 0.792 0.503 0.357 0.S45 0228 He rati ue design The designed distinguisher is in the form of a gate circuit (layers of simple interconnected functions]. It processes binary data and outputs a randomness verdict It is improved iteratively, using ideas from evolutionary algorithms (see the next section for details]. 5. Mutation & crossouer To form new individuals, we use makes small random changes in nodes and connectors. Crossover creates an offspring by combining different parts from two circuits taken from the population. The new population now enters the evolution cycle again, gradually improving its fitness. k. Suruiual ofthefittest Unfit individuals are discarded, better ones survive to the next generation. The higher the fitness, the bigger is the chance of survival. The evolution works as a heuristics looking for better individals - gate circuits distinguishing random and non-random data with higher probability than random guessing. Same graphic style should applied on each element Comparison to existing tools EACirc us statistical testing C\ (EACirc ^ ✓ Further information The standard way to asses jse batteries of statistical te= Dieharder or TestUOl. We n EACirc and c randomness is to s such as wsrsre, n them along with iinpare the results. To have a fine-grained comparison, we have analyzed 77 different functions (eStreom, SHA-3 and CAESAR candidates]. For 2-round Hermes and 1-round Fubukiwe confidently surpass NISTSTS. EACirc L.NIST-».* Diehard(er) f Interested in EACirc? See the papers referenced below or ask directly at the lab (CRoCS @ Fl MU). stinguishers foreStrewn nutí: >I.456[SECRYPT 2013). Sprir CRaCSě et GAP 2 02/11/0422. Case study Unity different styles of graphics Real-time Analysis of NetFlow Data for Generating ^ms% Network Traffic Statistics using Apache Spark ^^k\ Milan Čermák, Tomáš Jirsik, Martin Laštovička Institue of Computer Science, Masaryk University Botanická 68a, 602 00, Brno, Czech Republic E-mail: {cermak, jirsik, lastovicka}@ics.muni.cz Abstract — We present a framework for the realtime generation of network traffic statistics on Apache Spark Streaming, a modem distributed stream processing system. Our previous results [1] showed that stream processing systems provide enough throughput to process a large volume of NetFlow data and hence they are suitable for network traffic monitoring. This demo describes the integration of Apache Spark Streaming into a current network monitoring architecture. We prove that It Is possible to Implement the same basic methods for NetFlow data analysis In the stream processing framework as in the traditional ones. Moreover, our stream processing Implementation discovers new information which Is not available when using traditional network monitoring approaches. V_J Systems Performance Benchmark — Four Nodes (32 vCPUs) ■ Samza and Spark have a high-enough flow throughput and can be used for the analysis of data from multiple networks at the same time. ■ Apache Spark system has been chosen as it offers an easy management and a high versatility In terms of the running environment and proprietary processing methods (e.g., sliding window). Framework Architecture Row Collector The demonstration cluster consists of 7 virtual machines, one is dedicated to ipfixcoI, 5 to Spark and one to the Klbana and Web server. The following configuration is the same for all machines: ■ (4 vCPUs) lntel(RJ Xeon(R) CPU E5-2680 0 @ 2.70GHz, ■ 8GB 1600MMHZ DIMM DRAM EDO, ■ 85GB SCSI Disk with 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI, ■ 10Gbit/s network connection, 1 Gbit/s virtual NICs. Web Interface IPFIXcol Is a flexible IPFIX flow data collector designed to be easily I extensible by pluglns. In our demonstration, we use only part of Its wide functionality - data acquisition from multiple network probes and their transformation into a json data stream. itlstlcs Volatility of Network Traffic Data I Stream processing provides more accurate statistics about network traffic. m per 5 second interval volatility in results compared to traditional flow data processing approaches. ■ Allows to observe short, but , strong bursts of the network traffic that were lost due to the aggregation used In traditional batch approaches. Real-time TOP K Statistics Provides a real-time computation of Top K statistics to enhance network situational overview. mo í r Security Acknowledgements . ] M. Čermák, D. Tovarňák, M. Laštovička, and P. Čeleda. A Performance Benchmark for NetFlow Data Analysis on Distributed Stream Processing Systems. In Proceedings of NOMS, 2016. This research was supported by the Technology Agency of the Czech Republic under No. TA04010062 Technology for processing and analysis of network data In big data concept. Same graphic style should applied on each element Case study Grid grid MASARYK UNIVERSITY EACirc Using genetics to improue encryption Martin Ukrap, Petr Švenda, Marek Sýs, Václav Matyáš et alii Problem statement CRäCSe et GAP 2 02/11/04 22. Keeping elements within the grid Case study Unity # Connection & Disconnection Connection and Disconnection Connection and Disconnection Colour Elements with the same colours are perceiued as one group connection disconnection GE5TALT principles Connection and Disconnection Positioning Closer elements relate together one group separete shapes GESTALT principles Connection and Disconnection Shape Same shapes are perceiued as a group one group different shapes GE5TALT principles CERIT Scientific Cloud Looking for Synergies in Scientific Computing Connection and Disconnection Case study disconnection disconnection Mission CERIT Sdetffic Cloud centre, the successor of Supercompu-ting Centre Brno at Masaryk University, Is a national centre providing flexible computational and storage capacities. Provision of these resources is complemented with extensive research activities, carried both In cooperation with the user communities and in the e-lnfrastructure area Itself. History Supercomputing Centre Brno (SCB) is a part of Institute of Compt. Science, Masaryk University. SCB was founded mputing centres in the Czech :. Similar cooperating centres were Republic of that tim founded by other universities (Prague, Pilsen, Brni Ostrava). SCB has been working with Faculty of Informatics, Masary University, for a long time. The cooperation is both perse nal and factual, formally expressed, e.g., in a common re search intent "Highly parallel and distributed computatio Funding Transformation of SCB nto CERIT-SC will be s pported by a project of the 3rd axi of the RD&I Operatio al Program- me. The project will be realised from May 201 to October 2013. Its overal budget is 5 MEur. CERIT-SC is incl ded in he Roadmap for Large Research, Development and In novation Infra struct res in the Czech Republic. Scientific director Prof. RNDr. Luděk Matýska, CSc. Project manager Roman Čermák, M.Se, MBA http://www.cerit-sc.cz Goals CERIT-SC will provide highly flexible computation t rage capacities for the national e-lnfrastructure. Research and development in CERIT-SC is focused on ■ work with the users on tools and means for efficient use of the e-infrastru ■ cooperation with the users in development, deploymenl ■ new and modified systems and programs running ii jnd operation of flexible computatio ■ systems for storing, a ■tools and protocols for data storage facil The research work will evolve in a doctoral school with student participation from both IT and application areas. CERIT-SC will become an important node of national e-lnfrastructure, including integration into the European Grid Infrastrucute. This will be achieved by tight cooperation with CESNET on development and adoption of appropriate standards. Cooperation with Users e. Many existing ci Deluge of experimental data is expected in thods will break or stop scaling, new developments i User communities will come up with interesting problems, CERIT-SC w cessary IT expertise. We expect formation of joint teams ■ consisting of experts from both sides, ■ addressing specific research areas - both ad-hoc and long term work, ■ involving students (undergraduate and Ph.D.). This work will result in common publications. Targeted projects are alsi Formal agreement on future collaboration (Lol): ■ R&DI: AdMaS, BIOCEV, CEITEC, CzechGlobe, RECAMO ■ cooperating institutions: IBA,MZK, Loschmidt Labs., RECETOX ■ ESFRI projects (in negotiation): LINDAT-CLARIN, Euro-Biolmaging Flexible Resources Provision of the resources will range from traditional batch queues, through ii access upto the cloud paradigm. The resources will be provided free of charge. Prioritization of the users will be based on their scientific results; resulting ri cation will be achieved by technical means, combining advanced resource scheduling, vir-tualization, and the cloud paradigm; no complex administrative process will be required. By careful ballancing the scheduling strategies, successful users will get better share while new users, students etc. will not be prevented from using the resources. CERIT-SC computational resources are intended to serve unexpected and unplanned requirements of the users primarily. Data resources will serve to store and share data semipermanently and permanently. They will be tightly integrated with the computational resources. The target community are the end-users again. Equipment and Purchase Schedule The project will purchase the following resources: ■ SMP - Symmetric Multiprocessing clusters, with more t 64 cores and 128 GB memory per node (1000 cores tot; ■ HD - High Density clusters with higher number of nods 8-16 cores and 16-32 GB memory (2500 cores total) ■ HSM Hierarchical Storage Management (3 PB) ■diskstorage(600TB) ■ development tools and application software Q3/2011 Q4/2011 Q1/2012 Q2/2012 Q3/2012 Q4/2012 Q1/2013 cerit1" INVESTMENTS IN EDUCATION DEVELOPMENT Usage poster Position of the poster Ways of reading Ways of reading from Left to right Nauigation Position of elements, contrast, style of shapes could nauigate user how to read the poster. Input element The element which should be read the first. Colours and mouents attracts attention Ways of reading from Left to right Latin - from left to right Ways of reading from left to right Ways of reading from Left to right Ways of reading from Left to right Ways of reading Case study MASARYK jj UNIVERSITY Using ge Martin Ukrq| B je encryption r, Václav Matyáš et alii Problem statement Randomness testing The ciphertext produced by encryption should be completely indistinguishable from random data. But how to compare? EACirc is a framework for designing a distinguisher - a simple program that decides whether generated ciphertext looks random enough. EACirc workflow 1. Forming a population A set of currently considered partial solutions (gate circuits distinguishing cipher data from random data]. The initial population is created randomly. He rati ue design The designed distinguisher is in the form of a gate circuit (layers of simple interconnected functions]. It processes binary data and outputs a randomness verdict It is improved iteratively, using ideas from evolutionary algorithms (see the next section for details]. Z. Test uector generation r\ Testing data for learning is sampled from both sources. That is, non-random data from the C&~i inspected cipher and random data from a truly random source. 3. Fitness assessment Each circuit from the population is evaluated on all test vectors from the current set. Based on the outputs, it is assigned a fitness value from the interval [0,1]. "1—J—I—I-1 • ♦ ♦ ♦ i the evolution cycle again, gradually improving its fitness. k. Suruiual ofthefittest Unfit individuals are discarded, better ones survive to the next generation. The higher the fitness, the bigger is the chance of survival. The evolution works as a heuristics looking for better individals - gate circuits distinguishing random and non-random data with higher probability than random guessing. l. pictures - infographics 2. title 3. text Comparison to existing tools EACirc us statistical testing C\ (EACirc ^ ✓ Further information The standard way to asses jse batteries of statistical te= Dieharder or TestUOl. We n EACirc and c randomness is to s such as wsrsre, n them along with iinpare the results. To have a fine-grained comparison, we have analyzed 77 different functions (eStreom, SHA-3 and CAESAR candidates]. For 2-round Hermes and 1-round Fubukiwe confidently surpass NISTSTS. EACirc L.NIST-».* Diehard(er) f Interested in EACirc? See the papers referenced below or ask directly at the lab (CRoCS @ Fl MU). stinguishers foreStrewn nutí: )I.456[SECR¥PT 2013). Sprir CRaCSě et G AP 2 02/11/0422. Ways of reading Case study 1 FACULTY OF INFORMATICS Muuyk Urimnily Verifical B ogramswith Inputs ■Vladimir Still, Petr Ročkai, Jan Mrázek and Jiří Barnat DIVINE n DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library1 it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). A big downside of the current version of DIVINE is that for programs with inputs, this input has to be simulated by nondeterministic choice which is very inefficient. Therefore we present an approach for symbolic representation of inputs in DIVINE. .ibolic States consider a simple program with 32 bit input variable x and a branch on the value of this variable. In the current DIVINE, this program gives rise to 2iz possible memory configurations. In symbolic version, possible values of variables x and b are represented symbolically using bitvector formulae, therefore, there are only two possible configurations at the end of the program. Explicit Approach »Q2£j-»4^>-»| Proposed Approach SMT Solver Verification algorithm Interpreter 1 Correct ► Error To take advantage of symbolic representation of states, we transform the LLVM bit- in detail later. Apart from that, the verification al codein such a way that it represents variables which can contain values dependent on symbolic states with the help of an SMT solver inputs symbolically. This transformation is performed by LART and is resented Our approach aims for minimizing changes to the LLVM interpreter that is used to execute instructions in DIVINE The reason is that the interpreter is complex and performance tuned and therefore it is not desirable to make it even more complex by adding symbolic data manipulation into it. Instead, symbolic data are to be handled by the program itself. To encode symbolic manipulations into the program we transform the LLVM bitcode produced by the compiler and create symbolic LLVM from it. This not only minimizes changes to the interpreter, but the transformation can also be used for ■ data quite easily. The transformation is har Refinement Tool. Furthermore, DIVINE'S modified. It has to check if symbolic stat. they can represent at least one concrete state of symbolic states. For both of these tasks, DlVh from the program state and use SMT solver. Details of Program Transformation ilärt n- I LLVM j LLVM bitcode is generated from C++ source code. Dependen instructior control floi e graph of LLVM ; is created from the 'of a program. Instructio the input t s dependent oi _e computed. Dependent instructions are subs tuted with symbolic calls, path co dition manipulations are added. LART takes the LLVM bitcode of the program and libraries produced by the compiler and transforms it into a bitcode which manipulates data symbolically. In this modified program, any variable which can depend on an input value is represented symbolically using bitvector formulae. Bitvector formulae describe integers of fixed bit width with overflow and bitwise operations, and therefore nd therefore bohc variables rig ParaDiSe Parallel fc Distributed I_LIU Systems Laboratory are well suited for exact representation of computer integers. All the manipulations with such variables have to be transformed to their symbolic versions which modify the formulae accordingly. Furthermore, any branch which depends on an input value has to put constraints on the possible values of symbolic variables (this constraint is given in the form of a path condition formula). l. pictures - infographics 2. title 3. text loud Mies in Scientific Computing Ways of reading Case study Mission CERn Sctetlfic Cloud centre, the successor of Supercompu-ting Centre Brno at Masaryk University, is a national centre providing flexible computational and storage capacities. Provision of these resources is complemented with extensive research activities, carried both in cooperation with the user communities and In the e-lnfrastructure area Itself. History Supercomputing Centre Brno (SCB| is a part of Institute of Compt. Science, Masaryk University. SCB was founded in 1994 as one of big supercomputing centres in the Czech Republic of that time. Similar cooperating centres were founded by other universities (Prague, Pilsen, Brno, Ostrava). SCB has been working with Faculty of Informatics, Masaryk University, for a long time. The cooperation is both personal and factual, formally expressed, e.g., in a common research intent "Highly parallel and distributed computation CERIT-SC will provide highly flexible computation environment and primary data si rage capacities for the national e-lnfrastructure. Research and development in CERIT-SC is focused on ■ work with the users on tools and means for efficient use of the e- ■ cooperation with the users in development, deployment, and operation of ■ new and modified systems and programs running in flexible computation environ- ■ systems for storing, archiving, and retrieval of data ■ tools and protocols for data storage facilities interconnection, ... The research work will evolve in a doctoral school with student participation from both IT and application areas. CERIT-SC will become an important node of national e-lnfrastructure, including integration into the European Grid Infrastrucute. This will be achieved by tight cooperation with CESNET on development and adoption of appropriate standards. Funding Transformation of SCB into CERIT-SC will be supported by a project of the 3rd axis of the RD&I Operational Programme. The project will be realised from May 2011 to October 2013. Its overall budget is 5 MEur. CERIT-SC is included in the Roadmapfor Large Research, Development and Innovation Infrastructures in the Czech Republic. Cooperation with Users Deluge of experimental data is expected in near future. Many existing cc thods will break or stop scaling, new developments will be required. User communities will come up with interesting problems, CERIT-SC w cessary it expertise. We expect formation of joint teams ■ consisting of experts from both sides, ■ addressing specific research areas - both ad-hoc and lo' ■ involving students (undergraduate and Ph.D.). This work will result in common publications. Targete m Formal agreement on future collaboration (Lol): I K ■ R&DI: AdMaS, BIOCEV, CEITEC, CzechGlobe, RECAMO B ^ ■ cooperating institutions: IBA, MZK, Loschmidt Labs., RE ■ ESFRI projects (in negotiation): LINDAT-CLARIN, Euro-Bioli. íntific director Prof. RNDr. Luděk Ma Ly^no, ^:>c 'roject manager Roman Čermák, M.Sc, MBA http://www .ce r it -5 c. cz , Flexible Resources 1 Provision of the resources will range from traditional batch queues, through interactive ■ access upto the cloud paradigm. The resources will be provided free of charge. I Prioritization ofthe users will be based on their scientific results; resulting resource allo-i cation will be achieved by technical means, combining advanced resource scheduling, vir-tualization, and the cloud paradigm; no complex administrative process will be required. By careful ballancing the scheduling strategies, successful users will get better share while new users, students etc. will not be prevented from using the resources. CERIT-SC computational resources are intended to serve unexpected and unplanned requirements ofthe users primarily. Data resources will serve to store and share data semipermanently and permanently. They will be tightly integrated with the computational resources. The target community are the end-users again. Equipment and Purchase Schedule The project will purchase the following resources: ■ SMP - Symmetric Multiprocessing clusters, with more than 64 cores and 128 GB memory per node (1000 cores total) ■ HD - High Density clusters with higher number of nodes with 8-16 cores and 16-32 GB memory (2500 cores total) ■ HSM Hierarchical Storage Management (3 PB) ■ disk storage (600 TB) ■ development tools and application software 03/2011 04/2011 /2012 02/2012 03/2012 04/2012 01/ cerit IW) Function of image 1. pictures - infographics 2. title 3. text Ways of reading from right to Left Arabic - from right to left Ways of reading from top to bottom Chinese - ideograms from top to bottom Ways of reading Visual direction Location of elements The uisual weight of an element attracts neighboring elements, imparting direction to them Shape of element The shapes of an object creates an axis that imparts directional forces in two opposing directions along that axis Subject matter of an element Objects in a design may naturally point in a direction. For example an arrow. Objects opposing the intrinsic directional forces of an object can impart uisual direction to other elements in the composition Visual elements and Hierarchy LA5 VEGAS 5US MONDAY FEEMJAHT H.iOtl NEWS | 3 Approximately one in 40 area residents—some of wham may be your friends, colleagues or even your romatic partner - an looking for no-strings* attached relationships with people who are not their spouses. INFIDELITY CAN BE A CATALYST FOR CHANGE. IT CAN START A CONVERSATION. CNEATEHS FROM PACE 1: tlul ■ in 5 dram peLltuiu il\ol In lb |m>! )*M iu™d Ihr MdU net wn rfci n£ «iw » •(■dor Btdcrmin rewgniwd that in _rM*li>an,<*Uk:bI,l HfcuwUMih^cJonajg pink i mi purple- Hierarchy HIERARCHY: visual elemenets composed in a logical sequence crucial elements in contrast with elements with less importance layering of elements according to their importance position of elements leads the way how the image is to be read TYPOGRAPHIC HIERARCHY highlighting various information and its importance working with different size of font working with various weights of font Hierarchy 44 2626 prtxUn.i>>;iili.l m.Tvit.ť-vt.tHiip.inii-*- luve a tfnjiij» iiiipiK'i tm uurNK-ii'ty ;mil (.'iivimnniciit - mul Irmkc Qfl tntr qiulhy oflifc.hoth now and in the Hiuirx-. ili.n jCRaic* :L iXJtttly dliiy H)ľf;ifLvSj hi>w dc> v. .i: Lht up m ilim rcs]5t)iistliiliiy tiny in, ifclV ■'•Iii' ■■ ■ Rjtnp|iNr*m i in ipiMI» k> «H*M^M rvbpi lM>M*«f>4Mlll Hierarchy Hierarchy Hierarchy What's in here thin will be (really ) useful to you? Ogiluy's way of reading posters HOW WE LOOK AT A POSTER OGILVY: illustration > upper element title > above the illustration text > above the title Lemon. Ogiluy Think small. Olir -Alls carlsn'l KJ mxizt ol S (Ordrv firn vif nam. len i ' . . ■:■ ; a\ :: -■ j, Jr;n'r 0 l■ Ire jss ewlirji rJrjwn'i art O uieie i»* ubi Iiom. g N&b :■, iv-mi ' ■ jM* BtM li tc;i,U«t*pKinlB Wic-ivr'acMrlitlli Tt grjng nr'/gigs! rjjivi, mil' .n««3in5 cnli Ir^i-je. m^i^ »p WJO n)o! an ii lit of [ft ttMM 41li» .;0u gqt upjp lc dcarjl iMHFl yd! iifuMIa in'a a nal <".-» Or pcy n npcdr^tl Oi trsrJQ 1« VDir rjltf VYY +>■ a # Typography Typography Font style Font family Font weights Typography Setting of text boxes DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). A big downside of the current version of DIVINE is that for programs with inputs, this input has to be simulated by nondeterministic choice which is very inefficient. Therefore we present an approach for symbolic representation of inputs in DIVINE. left alignment DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). A big downside of the current version of DIVINE is that for programs with inputs, this input has to be simulated by nondeterministic choice which is very inefficient. Therefore we present an approach for symbolic representation of inputs in DIVINE. left justify alignment Typography Setting of text boxes DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). A big downside of the current version of DIVINE is that for programs with inputs, this input has to be simulated by nondeterministic choice which is very inefficient. Therefore we present an approach for symbolic representation of inputs in DIVINE. DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). A big downside of the current version of DIVINE is that for programs with inputs, this input has to be simulated by nondeterministic choice which is very inefficient. Therefore we present an approach for symbolic representation of inputs in DIVINE alignment width of a text box font size leading Case study Typography left alignment MASARYK UNIVERSITY EACirc Using genetics to improue encryption Martin Ukrap, Petr Švenda, Marek Sýs, Václav Matyáš et alii Problem statement Randomness testing The ciphertext produced by encryption should be completely indistinguishable from random data. But how to compare? EACirc is a framework for designing a distinguisher - a simple program that decides whether generated ciphertext looks random enough. EACirc workflow 1. Forming a population A set of currently considered partial solutions (gate circuits distinguishing cipher data from random data]. The initial population is created randomly. Z. Test uector generation ^ Testing data for learning is 5 sampled from both sources. That is, non-random data from the £"A") inspected cipher and random data tSjl i from a truly random source. " -i & * v t 3. Fitness assessment Each circuit from the population is evaluated on all test vectors from the current set. Based on the outputs, it is assigned a fitness value from the interval [0,1]. "1-J-1-1-1 • ♦ ♦ ♦ i 0.792 0.503 0.357 0.S45 0228 He rati ue design The designed distinguisher is in the form of a gate circuit (layers of simple interconnected functions]. It processes binary data and outputs a randomness verdict It is improved iteratively, using ideas from evolutionary algorithms (see the next section for details]. 5. Mutation & crossouer To form new individuals, we use makes small random changes in nodes and connectors. Crossover creates an offspring by combining different parts from two circuits taken from the population. The new population now enters the evolution cycle again, gradually improving its fitness. k. Suruiual ofthefittest Unfit individuals are discarded, better ones survive to the next generation. The higher the fitness, the bigger is the chance of survival. The evolution works as a heuristics looking for better individals - gate circuits distinguishing random and non-random data with higher probability than random guessing. Left alignment keeps same gaps between words Comparison to existing tools EACirc us statistical testing C\ (EACirc ^ ✓ Further information The standard way to asses jse batteries of statistical te= Dieharder or TestUOl. We n EACirc and c randomness is to s such as wsrsre, n them along with iinpare the results. To have a fine-grained comparison, we have analyzed 77 different functions (eStreom, SHA-3 and CAESAR candidates]. For 2-round Hermes and 1-round Fubukiwe confidently surpass NISTSTS. r*-m EACirc L.NIST-».* Ba-m-BsL^ Diehard(er) bb^. f Interested in EACirc? See the papers referenced below or ask directly at the lab (CRoCS @ Fl MU). stmguislwr; foreStrewn nutí: )I.456[SECRYPT 2013). Sprir CRaCSě et GAP 2 02/11/0422. Text and legibility Black text on a white background allows common speed reading. White text on black background reduces reading process of 15 %. Text and legibility Black text on a white background allows common speed of reading. Text and legibility White text on black background is optically thicker. legibility Text in Italics reduces redabiLity of15 Text and legibility LONGER UPPERCASE TEXT REDUCES REDABILITY OF 15 % Text and legibility TEXT Left alignment doesn't haue any affect on speed of reading. Left alignment doesn't haue any affect on speed of reading. TEXT Left justify alignment doesn't haue any affect on speed of reading. Left justify alignment doesn't haue any affect on speed of reading. • Typography and errors A TYPOGRAPHIC RIVERS D r. Jozef Ferenczy učil celé generácie študentov a nitrianske gymnázium malo aj jeho zásluhou veľmi dobré meno. Okrem Karola Pongrácza k jeho žiakom patril okrem iných i maliar Maximilián Schurmann (1890 -1960). gaps in typesetting, which appear to run through a paragraph of text, due to a coincidental alignment of spaces spaces caused by full text justification or monospaced fonts no hyphenation Case study Typography left justify alignment Systems Performance Benchmark — Four Nodes (32 vCPUs) Samza and Spark have a high-enough flow throughput and can be used for the analysis of data from multiple networks at the same time. Storm I Spark I Samza I Ml! i + Elasticsearch Apache Spark system-has been chosen as^t offers! an easy management^ and § a high versatility in terms of the running environment and proprietary processing methods (e.g., sliding window). Left alignment keeps same gaps between words The demonstration cluster consists of 7 virtual machines, one is dedicated to IPFIXcol, 5 to Spark and one to the Kibana and Web server. The following configuration is the same for all machines: ■ (4 vCPUs) Intel(R) Xeon(R) CPU E5-2680 0 @ 2.70GHz, ■ 8GB 1600MMHZ DIMM DRAM EDO, ■ 85GB SCSI Disk with 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI, ■ 10 Gbit/s network connection, 1 Gbit/s virtual NICs. < nterface IPFIXcol is a flexible IPFIX flow data collector designed to be easily extensible by plugins. In our demonstration, we use only part of its wide functionality - data acquisition from multiple network probes and their transformation into a JSON data stream. _ Real-time TOP K Statistics Provides a real-time computation of Top K statistics to enhance network situational _ • 80 (http) • 23 (telnet) • 993 (imaps) • 22(SBh) • 143 (imep2) • 445 (microsoft-ds) • 25 (smtp) • H0(pop3) typographic riuers without hyphenation Infographics Infographics THE CITY DEFINES FOUR TYPES OF VENDORS FOOD VENDORS GENERAL VENDORS FIRST AMENDMENT VENDORS VETERAN VENDORS UNLICENSED VENDORS f * i® ¥ E3?=S5E. Hit MttM T^7 tttttrnttf* MtttttttMt ULC-o J® Infographics A «lectrfcJdadfe cm Portugal De onde vent a nossa luz? A sletricidade quechega as nossas casas vem de umcomplexo cabas de tonces energéticas. que vana conforme □ ano. Em 2008, houve muiia importagäo e pouca producSo hidroeléc trica Mas a energia do vento já corner a ter um peso sigriiftcötlvo RiCdTdaGdrcid ŕtexŕoj Q Jaaquim GveitelFS finfogľšfid} * ÄS remuV j Ye Li ■ •■ . ■ a auäfigäl. in Ji liirtnifií ainiii ilupmJe i rroJiiMr a (ľtcítrlcl d a de J e 1411 í rsoíisila. hm jihu;, ys tcrawriícm- _ jc-nw rating ju iiiejuciid. Be j.Lidu íiíiilIu riasu ľimiiiiIľí tliuHilĹH-ailai. ba rr-niiř 4 eilv -r-srti val eo paíí dt|i;rKíia fonaiKtiH! do c^rvao, ujuc isníítäo [inha ahorintiiKi dríiaiíLS uuum;ií(k]ĽlLi-lrfLUnlu. dailu : iiíľ i ií:I ii'.i -.-f\ jji i . 11111 -_- iT! ■_■ i ■ -i■ nii Ľtam ä •■.i. j i ■ M p-mh^in i*jrtrii=i ■ - -■ ■ 1 vflriinäw. Ij wnlw. iííaitii [kiheuJii. [uriiL-nsu quasi! tahá Ěůť+fjia l|Iľitíu hjLuríu^Liw, ruduřSiiJo o děéo daj m n« t raní f qiKi, wlm lf efititÄ JD dimnii[lis"íKiias" rnuiiírcis • i i. i ainila apMla* marfinaJ-rtiLTiKr)unn>ttflĽi linkami iir.'.ľ.iid.ul.i.ii :d,: i-.-k L.iir iu quŕtn dlwwHj-polfaĽolľKBdĎ. \ proJinJi-i t■■ --j i----1-1 de- |jiimiisřiiliuririEiJln/nlíiiÍLin. rm infaiuru liní furtínlĽ. A pán.tli milí íAllludlrtOriHi etuOSleldildiío i. :.l.i.|l ľ|i i..!. rin, p-slses. j>iSPCOES5Íd«!eí de unpmta í») vjíiwi unoiiiiiKJ.-MJrjrniniĽ tí f™jiiľ i .i-...i.I . d 1.1.1. .-.i I ■ dl l'ijfrioj.il m ]ü j»r ífimB di otot m d dad o u^w TOtiŕijjiiln, 4 vneríia •kk de viiippTO.tĽimc E*p»riha. u,ijr Ivni yulm fiihin eiipcj}ěti^. iinrte c*aýn hi. I'M.Li.iiii.i.^ ukVirtiĽaS. .'.n. i.iiíi 3 ■ -i 11. .-IľUli •■■n. i ■■.. i .1 ism d i1 a&i srKUrnodit ^umi am uu rjSií,w.pompKS« oiriKOTTKrai rmiiBÍ-n nuclmc. - Ill i...........il L ' ' iii|iiiliii- I v," — 11 r *■ I-lIiIií i Hrlŕi4nhqu liJili/. illjbjni E-OUHCE: USD*, Economit Research Service, Feet) to inj DataBase; Benewaoie Fuels Association. # Process of creating poster Define a grid Define a layout Define a positon of text and pictures Hierarchy of information 2 Compose elements Try to find balance between text an images DIVINE DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). n I LLVM I FACULTY OF INFORMATICS Masaryk University 3 Typography Choose typography Find a way how to highlight the text - working with different styles (italic, bold,...) DIVINE DIVINE DIVINE DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). DIVINE is a tool for verification of parallel C++ programs. By using the LLVM compilation framework with the Clang compiler and the libc++ library it provides support for most of the standard C++ library and all the C++ language features. DIVINE is rather efficient when dealing with programs without inputs (for example test cases). Case study Good balance between pictures and text A eicctTicidade cm Portugal How Visual Weight and Direction Impact Design Balance Your composition needs to be in balance, whether symmetrical, asymmetrical, or radial. You'll achieue this balance by placing elements of combined equal uisual weight on either side of the optical center Dominance/Focal Points Focal points are elements that attract the eye. They're elements of greater uisual weight. The dominant element of a design is the element with the greatest uisual weight. Flow Through focal points, hierarchy, and uisual direction you can lead the eye from one part of your design to the next. You'll create a flow through your design. Scale Is generally considered to be the relative size of different objects. Here we can consider it in the context of the relatiue uisual weight of different objects. Depth Elements with greater uisual weight appear to moue forward in a design while uisually lighter elements recede into the background. We can use this understanding to create depth in a design. Proportion Is the relationship in scale between elements. Different proportions in a composition relate to different kinds of balance and can help establish uisual weight and depth. Hierarchy By creating a scale of focal points or elements of different uisual weights you can create a hierarchy of design elements. The difference in uisual weights is what makes certain elements stand out improuing scanability. Case study POSTERS Case study Posters Case study Posters Ultrasound Painting of Vascular Tree text-title Asmund fiirhvta ► text images ► logos Case study Posters Ultrasound Painting of Vascular Tree Suggestions Uoivenit\ of Ultrasound Painting Vessel Tracking Dub is tn« high, haw tale ul iflltawui"). Uie >< ■mw to HMrHap I .i-i' ■ ■ .nf» (rain* a navr wed-fmlnl h «ainiltteil «i the ..wiiei CUI1I (raamtrananritflPdai an educated at wtwr n ■ em I «ii. iF-.- n«rt hin* To acted bfanthfiig. we calculated additional Mttl intnH outet «je01 ih«c«M t«t»ot>. finwhw^ on i«i *i it* *r« oeinh In the Mtetrton' at a ■ ■ dm Ihm a rvgutm rtiilrthWion or t«tfnt, an the ouUMr Ii iiHM 1Mb a ip!» lift «IHl tl« »I K »l>lt«" Mfl « Point Cloud Processing ««♦■«frari «rw im an«, tiw iriaugglaitai MF bawd on win! iwrmal and Hip I IMEDVIZJ ILLUSTRA SOUND,,, Case study Posters Ultrasound Painting of Vascular Tree . md Girkeland and l>an Wla Ultrasound Painting Suggestions ^ more space between elements Case study Posters Ultrasound Painting of Vascular Tree . Suggestions JUinund 8tt*rtand and Kan Viola Ultrasound Painting U ts* **■ Mr. MA) Vessel Tracking > Qua to th* high trante*ateul u1lramii<4, Ihevi A n*w wl n™i tl rtlniltttil At I h* ■ <■••<<> mil firm swuvii Mid inr.1 n in educated pun) at whpir il>--«csm« will b« m (h« nctt fum* To (KIM Na.rw.mnj:. vw .-al..uUI«i1 .^Minimal Wftl lidlnit in iXllpt Wjeol ItWO&rt BTancMrjotiim «1 the tM-ffoMih hi llw 'Mwlptwi" of i ■rtw i* i a iwfluLar diittnuiKm r.r iwinti on itw miHinr i» ml-' a H3 (mM rrtautf ami ihr wtfpfHdDeAil imKIk! ■■iii> a iw more space between elements w background photo makes text illegible Case study Posters Ultrasound Painting of Vascular Tree . JUinund 8tt*rtand and Kan Viola Ultrasound Painting Ihr euHtn* •kam MM Vessel Tracking Out us ir»hi|hframe lalsrul irtli««uu» swd-öomi h ■ =■.■■>! -i'f-- Dt th» ■■-<•■■' mm fKltan *rd utKi »i an cduc«tM) punt atwfwiv il>--•rt« I* m 0» rxirt (tin*. To .Iclmi Nurwhuij:. vw enlculMMj n, Minimal Wfttinjlnh in duti* «feol ir«tn-.i p'-"1 h ■ =■.■■>! -i'f-- «I Uw ■■-<•■■' trwa Mlm Mid u^ai an educated iuw whw* ii>-- •rt« I* tri 0» ncrt Mil* To dated btwKhtn( v»v .-al.uuinl aiMiunnal i-vit i-nm- in dutt* Mfe of irwt-oM iww*. BWKMnt«rm #l the tfti-poMihtollw 'Mietptwi" of a a rvjutai iHiMWittcm .-r v■. ,■„■, the QtiUlne I) added Mils Point Cloud Processing •jOtfed :r.::,. I- Ifulaltoil -Ii ".I the I" ,ii Suggestions iw more space between elements ^ background photo makes text illegible ^ unite the size and positions of elements iw too many overlapping elements - uisualizations, photos, ... Case study Posters Ultrasound Painting of Vascular Tree . Asinund nrtttllMW and Kan viola Ultrasound Painting The .-. ■ r I i.t «I HB t ewnnnrM Vessel Tracking Out to irwhifriframe labeul iiHrawiii«!, Iii« w A rm seed-^orni hcalculated ai the «kirnt .■m« «viuv. Mid medal an educated pue»» atwheie il>-- To .Iclmi brant hm« vw calculated 8iMlllnnal teed pdlnh in i>uiet «je oi iiMt-w iww». BrarcMnt«rm #i the «M poinh la the "! lie Irion" cl a nwwiw a rvjulai iHiMWittcm -r >-■'. t■■ on the willlne h *Mni Mite Point Cloud Processing Mtded polnlt :<-■■.■.!. I-ifulailo" and the le**l Suggestions more space between elements ^ background photo makes text illegible unite the size and positions of elements too many overlapping elements - uisualizations, photos, ... more space between logos and graphic elements Case study Posters Ultrasound Painting of Vascular Tree . Asinund Strkeland and Kan Viola keep a safety zone around the logo Ultrasound Painting e*an>in*i Vessel Tracking Du» to irwhigliframe ratp-jl idtiawuH. M'c " .'•V'lüll Ii«« ':,lil'v n 11 J"'-" A n«m >n-.i p'-"1 h ■ =■.■■>! ■( th* mm iKlran *rd m»»l »1 an cduc»tM) punt a( «ihm« il>-- Tfl .Iclmi btwKhtn( v»v .-al..uUI«i1 a^UHIinnal m*l ndlnh in ixjin «djeol ilmiMviwwi. Branching. «nm ailht WO poMnh la II» "llwlriün" of a now dun a rvgutat nnlrmutKiri ,-r t ■. on the QtrUIIW *Mnl lillti Point Cloud Processing and ttw local Suggestions >w more space between elements ^ background photo makes text illegible unite the size and positions of elements >w too many ouerlapping elements - uisualizations, photos, ... ^ more space between logos and graphic elements Case study Posters Nis escienime namende corepe nobis ullam, aut elenda coris exero volorpo ressund empöre pa vendem. Bit ut aligendunt alique isti cusa nihicto tern sant quias ea dolorum eumquatem quae nulpa cuptata ipisciuntis dolorescic^et/abor alitinc imilleste (pro voluptata\ qui rectotatur, nis volendam, / ut consedit / aut ut porumleos samusae/ consequodi dunt excerio/consedit iderum ne sum /volorib usapediscias. Nisescienimenamendecorepenobisullam.autelenda coris exero volorpo ressund empöre pa vendem. Bit ut aligendunt alique isti cusa nihic to tern sant quias ea dolorum eumquatem quae nulpa cuptata ipisciuntis dolorescid et abor alitinc imilleste pro volendam, ut voluptata qui consedit aut rectotatur, nis ut porum eos samusae consequodi dunt excerio consedit iderum ne sum volorib usapediscia wider text box Typography nuers Case study Posters Nis escienime namende corepe nobis ullam, aut elenda coris exero volorpo ressund empöre pa vendem. Bit ut aligendunt alique isti cusa nihicto tem sant quias ea dolorum eumquatem quae nulpa cuptata ipisciuntis dolorescic^et/abor alitinc imilleste (pro voluptata\ qui rectotatur, nis volendam, / ut consedit / aut ut porumleos samusae/ consequodi dunt excerio/consedit iderum ne sum /volorib usapediscias. Nis escienime namende corepe nobis ullam, aut elenda coris exero volorpo ressund empöre pa vendem. Bit ut aligendunt alique isti cusa nihic to tem sant quias ea dolorum eumquatem quae nulpa cuptata ipisciuntis dolorescid et abor alitinc imilleste pro volendam, ut voluptata qui consedit aut rectotatur, nis ut porum eos samusae consequodi dunt excerio consedit iderum ne Settings of type and type box influences the layout nuers left alignment Case study Posters Nis escienime namende corepe nobis ullam, aut elenda coris exero volorpo ressund empöre pa vendem. Bit ut aligendunt alique isti cusa nihicto tem sant quias ea dolorum eumquatem quae nulpa cuptata ipisciuntis dolorescid^et/abor alitinc imilleste (pro voluptata\ qui rectotatur, nis volendam, / ut consedit / aut ut porumleos samusae/ consequodi dunt excerio/consedit iderum ne sum /volorib usapediscias. Nis escienime namende corepe nobis ullam, aut elenda coris exero volorpo ressund empöre pa vendem. Bit ut aligend-untalique isti cusa nihicto tem santquias ea dolorum eumquatem quae nulpa cuptata ipisciuntis dolorescid et abor alitinc imilleste pro volendam, ut voluptata qui consedit aut rectotatur, nis ut porum eos samusae consequodi duntexcerio conse-.dit iderum ne sum volorib usapediscias nuers smaller font size + hyphenation Settings of type and type box influences the layout Case study Posters I I L 1111 1 mlnlmumsavstanderi til andre grafiske elementer Logo safety zone THANK YOU :)