14

Click here to load reader

Computer-aided production of scientific documents

Embed Size (px)

Citation preview

Page 1: Computer-aided production of scientific documents

ComputerPhysicsCommunications61 (1990) 163—176 163North-Holland

Computer-aidedproductionof scientificdocuments

JosefBrandtInstitut für Organische Chemie, TechnischeUniversitdt Mdnchen, LichtenbergstraBe4, D-8046Garching Fed. Rep. Germany

Principlesof computer-aidedproductionof scientific documentsaredescribed.Thehistorical developmentof text-editingandformattingsystems,andthedifferencesbetweenthesesystemson onesideanddocument-producingsystemson theotherside are~elaboratedunder theaspectof immediateuserinteractionand logical structuringof documents.The “man—paperinterface”,i.e. considerationsof perceptionandacceptanceof printedvs. computer-readablemedia,areshortly discussed.

1. Writing as a tool of scientificwork • the man—paperinterface, i.e. presentation,re-cognition and perception,including the art of

Writing — in the broadestsenseof the word — t)rl)Ography.countsamongthe indispensabletools of scientificwork. Onemajorpurposeof writing is theproduc-tion of a document.The documentis an account 1.2. Historical developmentof the scientificwork, andthusrepresents— possi-bly accompaniedby samplesof materials,speci-mens of organisms, graphs, computer-readable At this point a reminderof the historical back-material, etc. — the communicableresultsof such ground is appropnate,becauseit has stronglyscientific work. influenced the presentways of thinking. Before

the adventof computers,anauthorwould producea manuscript that usually bore the traces of a

1.1. Researchanddevelopment tedious editing process: erasures,deletions, ad-

ditions, cuttingsandpastings.Hewould — usuallyWriting, or as we should more appropriately through a middle-manfrom the publishing trade

say: the principlesof productionof readabledocu- (a lector, editor, etc.) — hand this “final”ments, is not commonly regardedas a primary manuscriptto a representativeof thegraphicsartsobject of scientific research,but rather as one of who would then transformit into a typographicalits tools. Owing to the importanceof this tool, presentationandwould eventuallyproducean ap-thereis, however,a growing amount of genuine propriatenumberof copiesof that presentation.basicresearch— as distinct from developmentand As the communication between author andengineering — being done with regard to the typesetterwasusuallyrather indirect, the processfundamentalaspectsof the productionof docu- was slow and error-prone. And: it made laterments.It concernsquestionslike: alterations of the documentrather difficult and• logical structure of documents,including the expensive.Such alterations will, however, arise

aspectof standardisationandtransportation; rather frequently becausemanuscriptsare never• the man—machineinterface,i.e. the functional- really final. They only representthe statusof a

ity andtheusersurfaceof a document-produc- documentat agivenpoint in time (why elsewoulding system; neweditionsof abook differ from previousones?)

0010-4655/90/503.50© 1990 — ElsevierSciencePublishersB.V. (North-Holland)

Page 2: Computer-aided production of scientific documents

164 J. Brandt / Computer-aidedproductionofscientificdocuments

.,‘_ ~~4•’ -. .,

- ~ ~.. e-~ ~ . ç,~ i . i

~ ~

k~’

c~ W6fl ~ __/7”~ ~-~‘ ~ ~ ~

P~oiggesdclI*It*. oa dim grimi p~fr~qs p~.ali P~~ ~~er ,e~u~t d~cluttir emc I. — ~ —1~4,,~/-~ -.l.~! Lr~~~a.upa~ I fr’sZ~~tP-’

£~‘il~ñ~ ~_LIJ melguc. i~iI~~ mfl ~ /

~t 10 iuftandee petheesmdertha1...,~ / ,,dm tyranmeid~tail ge, .‘— •~-‘-~i~~ ~/__ ~‘ ~ lp~atIiptti elfriyer un ‘-~—~~‘~ ,, j,

~ ~ p~rfifli ~ ~~ ~~ ~

1~i1~±~ ~1/”) -ii.. r~p~-~~j~ -

/ ~tçl ________________ ?f} rVai~pr~Ii~.do I’i~ouim~ce employdopr

1d~dsiIt~4~S’ ~s~‘ CO lb toi quaI.. ‘‘,..

/,~ .*q~utàI’ _______ _~ ~••.d__~,nek ilee g ;r~-~ j bI.~/P ~~ ~ ~ ,,~. ~ ‘1 a

bi~.

II • mc .1. ~,qz~I”i , __________________________ ~. S./ ~ ar

—P7/ gale . /fr__a U4# ~fI d ~ua~.$LM~NJS

~Z’Z’)~‘~ ___~ lb ~ /~..d5 -

~ Ill bomans MS 1—4’oat at do Ia j fee~ $1~j. I

c’ av.nst Tf~(/(i~j,j4’ anne franca anal a scan, an ~‘

l’C . ‘I 47 ~

I iIL’~ - ) ~ .7

~t •~~::~

54,~ir _~a1s_..i~a,.

~ 7 .). .41. 4’J ~ /i ~

Fig. 1. A remarkableearlyexampleof “text-processing”[1,p. 97].

This resultedin a costly, sometimesiterative, pro- Honoréde Baizac’s famousgalley proof, repro-cessof proofreadingand re-typesetting.Figure 1 ducedin ref. [1, p. 97] (seealso ref. [2]).gives a remarkable example of this process:

Page 3: Computer-aided production of scientific documents

J. Brandi / Computer-aidedproductionofscientificdocwnents 165

1.2.1. Textediting “TYPESET” [4], “PROSE” (written in CDC-It is by no meanstoo remoteto rememberthat PASCAL in 1979) etc. Today’s quasi-standard

shortlyafterthe eraof toggle switchesandconsole WORDSTARalso belongsto this family ~.

lights,therewasthe periodof teletypewriterswith In the course of time, more functionality waspaper tapes, and of card-punching,sorting and gradually addedto such text formatting systems.readingequipmentwith punched-holepaperboard A remarkablestep forward was the addition ofcards. And with them came the generationof some sort of programmingfacility which cameinline-orientedtext editingprograms.They relieved the form of a macrolanguage.the author from the tediouswork of erasing,in- More and more commandsfor font changes,serting, cutting and pasting their paper manu- subscriptsand superscripts,proportionalspacing,scripts. etc. wereaddedto suchtext formatting software.

But they have engendereda way of thinking Thus, the representation“languages” were farthat still persists althoughthesemedia may have from standard,andportability a dreamof somevanished.(How else could one explain the fact optimists. This last aspectis by now drifting to-that thetext editor which came— andstill comes wards a more acceptablesituation with POST-— with today’s most widespreadpersonalcom- SCRIPT gaining increasing acceptance.Its de-puters’ operating system is a program called mandson computerresources,andto someextent,LINED, which is exactly what its namesays: a also, the licensing policy of ADAGE result inline-oriented text editor.And this is on amachine fairly high pricesfor POSTSCRIPTequipment,atwhich neverhad acard readeras its input, nor a leastat the time of this writing.typewriter as its consoleoutputdevice,but which, Terminals, in particular the ubiquitousCRT-from its very beginning, was conceivedwith a displays, were (and mostly still are) unable toCRT screen.) display thefinal appearanceof text in the printed

document, becausefor the sake of speed they1.2.2. Textformatting preferredto usetheir built-in charactergenerators.

The next step were text formatting systems The advantagesof WYSIWYG fell behind, andwhichbrokethe conceptualboundaryof a line of were/are mainly left for modest text editor/text being the logical basic entity of texts. They formatter programs,while with more ambitiousconsidereda pieceof text, usuallydelimitedby a software,the userwasfacedwith a multiplicity ofcontrol code,or blank lines, and calleda “para- control-key or function-keystrokeson input, andgraph” to be the fundamentalunit. Initially, they ratherconfusingcontrol codeson theCRT-screen.cameas separateprograms:the text itself, inter-spersedwith formatting commands,had to be 1.2.3.Desktop“publishing”preparedon a separatetext editor. With the technical developmentof printing

Most of these text formatting programs equipmentshifting from preformedcharactertyp-descendedfrom a small number of ancestors. ing toward pixel graphics,the long existing needAmong thesewasRUNOFF whichwasdeveloped for greatervariety in shape,size, positioningandon theDartmouthtime-sharingsystemin the1960s orientationof characterscouldbe fulfilled. Usersand becamequite popular on DEC’s PDP1O/20 suddenlyenjoyedthe freedomof designingtheirand PDP11 systems. Later, “RUNOFF” was owncharacters,logos,layouts,etc.,sometimeswithshortenedto “ROFF”. Usersof UNIX will recog- rather bizarre results. The “desktop publishing”nize RUNOFF’sshort namein nroff andtroff. programmerely presentsthe tools, but doesnot

A furtherstepwasthe theincorporationof text offer any help to their proper use. The mainformatting into the text editors themselves:thus, burdenof where to put what, in order to achievethe CALL-A-COMPUTER-System,e.g., provideda RUNOFF-version(called “EDIT RUNOFF”) *i . . . . .

Oneshibboleth of this family is theperiod-character( .

whichwasmcorporatedas a text editorcommand. in the first column of the input text which servesasthe

From these stemmed further offspring, like commandescapecharacter.

Page 4: Computer-aided production of scientific documents

166 J. Brandt / Computer-aidedproduction ofscientificdocuments

ABCDEFGHIJK ABCDEFGHIJKLMNOPQRSTU LMNOP~STVWXYZ abcdefUVWXYZabcdghijklmnopqrstu efgh~jk1mnopqrstvwxyz1234567890UVWX~Z1234567

~mbo

ABCDEFGHIJ ABCDEFCHIJKLKLMNOPQRSMNOPQQ1~SfUTUVWXYZab VWXYZabcdefgCdefgh~Jk1mnopqhijklmnopqrstuvrstuvwxyz 12345 wxyz 12345675906789O&P!f~$(;)~&~$~)~:Ba~kervtIIc [‘alatino

ABCDEFGHIJK ABCDEFGHIJKLMNOPQRSTU LMNOPQRSTUVWXYZ abcdef VWXYZabcdefghijklmnopqrst ghijklmnopqrstuuvwxyz123456vwxyz1234567789O&?!B~$q 89O&?!B~$(;)~Fig. 2. Someprofessionallydesignedfonts[10, fig. 110 a,b].

an optimal result, is still left to the humanoper- 1.2.4. Typesettingsystemsator. Theplethoraof professionallydesignedfonts As wehaveseen,“desktoppublishing” — as the(figs. 2 and 3) is not generallyavailable,eitherfor word is usednow — actually is amisnomer.“Pub-practical reasons(cost of digitalisation, storage lishing” implies much morethanjust the ability ofrequirements),or becausethe moreattractivefont placing characters(or any other shapes)at arbi-designsareprotectedby copyrightor expensiveto trary sites, positions, orientations,and perhapspurchase. colourson somesubstrate.One step of the tradi-

With the shift away from characterstowards tional processof publishing is typesetting. Thepixels, incorporation of non-charactermaterial, authormerelyproducesamanuscripttogetherwithi.e. graphics,becamefeasible,andimmediatelythe some general ideas about the appearanceof thequestionof a standardfor a graphicslanguage document.It is the publisherwho then providesarose.Here, too, the fog is clearing,and we ob- craftsmanshipin typography,proofreading,layout,servea move towardsPOSTSCRIPTfor rendition etc. “Desktop publishing” systemsdo not evenof what I may loosely classifyas “artwork”, and providethisjob of professionalexperttypesetting.towardsHPGL for vector orientedgraphics. It was therefore absolutely logical that the

Page 5: Computer-aided production of scientific documents

J. Brandt / computer-aidedproductionofscientificdocuments 167

N*$~D�11SR~Sikcdy~mE~OcOOIYfiPICD~eTwrite~ .• -‘

- ~r’~,w?i,i’i/,~~ c~EE~Sn. 8o~d

~i$7~~I~irn ~ “

• ~ M ~I• ~

4~ H

Fig. 3. A menagerieof decorativefonts [10,fig. 110ci.

problemof computer-guidedtypesettingwas iden- Let me illustrate by a trivial examplewhat Itified as an areaof researchin computersciences meanby logical structure: the title of a document[6,7]. Since this touches the present, I shall is a logical entity. Its physical appearancemayterminate the historical excursion at this point, vary due to extraneousinfluences,and is subjectand shall elaboratesome more on this aspectin to tasteand fashion (see fig. 4). The title of ansection4.3. article in, e.g. ComputerPhysics Communications,

may be typeset in a font, size, alignment andpositionthat is quite differentfrom onein a series

2. LOgiCal structureof documents of internal reportsof an author’sinstitute.A titlemaystartanewpage,it maycausetheinsertionof

One of the fundamentalaspectsof document ahorizontaldividing line or vignette (asin fig. 5),processing,as opposedto word or text processing, or just someblank space,possibly accompaniedis the conceptof the logical structureof docu- by a check for a minimum required amount ofments.Good typographyis onemeansof reflect- availablespaceon that page.ing this logical structure.

Page 6: Computer-aided production of scientific documents

168 J. Brandt / Computer-aidedproductionofscientificdocuments

SIMPL !CISSD4US

f~amrnV’aganrrn/ genant ~ttfcf~ioreumfr(~ton ~uas1~atm/ t~unt’ I~ei~rgi~t~rutmltcbInbkft~2*JUomnsm/ivo~rthngtfIbrn/gdinut/*~~$tiuunDat4.

~ftauDtu/aucbMnnnbrfolc lvkDcr

msq(id, 311 lefcn.un¶Za~4ebefl

sonG~is*~SCHLEIPHIIM

von SuIs(ort.

~cbmdt __

Fig. 4. A title pagein baroquestyle,usinga variety of fonts [3J.

2.1. Sections and paragraphs, .. . and the stuff title page, impressum,etc., each time and eacharoundthem country has developedsome rules of the trade.

Stylesfor the correspondingentity in journals(i.e.An obviouspieceof structureis apparentin the title, author(s),acknowledgements,abstracts,etc.)

“body” of a document,with its hierarchyof parts, are much more subject to variation, not only inchapters,sections, subsections,paragraphs,and typographicalstyle, but — more serious— in theirfinally sentences.Thesearemarkedby headlines beingpositionedasheadlines,or on separatepages,or titles, respectively,with the typographicalap- and — for acknowledgements— togetherwith thepearancereflecting their positionin the hierarchy, title materialor at the end of the document.Anand may or may not be separatedby blank lines “intelligent” systemhas to be smart enough toor pages,or line breaks,etc. representsuchvariationsby meansof its informa-

Also for the preliminariesof a book, which in tion aboutthetypesetting“style” of a given coun-German typesetters’ language has the name try, editor or journal.“Titelei”, meaning the ensembleof frontispiece,

Page 7: Computer-aided production of scientific documents

J. Brandt / Computer-aidedproductionof scientificdocuments 169

~)on~ont~af4,cün~9c~~cccten.~ vie~oev~tettm Contr~feL~tungcrntrifftifoli menigEIid,

~ tviff’rn / b69id~inbiclwmeinwavbeitvubevft4nbenL~ab/einerjcbCfl6t4tt / bermbefd~rcibunginbiefem ~ud,vev~frft if~/geLegen~eittrnb contvafebtifd,c pictur/fovzeI mug~

Iid~/eyn~u1eiben/~b~ud~beshalbenmid) mitfd~reiben/vub buvd,mittcl pcvfonen/weltmbbreitbcwovbrn/nicbt4UeininZcutfd)enz~L~ibe/fonber and) in ‘5taLia/j~rancheid)/£�ngeItrnlbt/~~oLanbtvt~enmnavc~.fl~aeId, aber crlangt L~abbey ettid~enffârften / ~fd)offcn/@tetten/vnb etLid)en belunb evenpcrlonen/wirb inbiefetn~iucbmit ewigem!ob beven/fo jbv bi1ffI~er~uge4~anian jebem~rtgemeibct.Don ina~id)cmort i~ mwauffmciii antangen ~thianttvortwort’en. !Eo t~atlid)aud~manid~ore be~tagt/ba~ce minnld)thm~m~gcn;un’ilLm wcvben/ eunc~gefd)ic~tentflatcva ~,a1b4~X~iebairnid~aud~bey ctlicbmgvoff~netcttcn erfa~ren1,ab / ba~nid)tehnje~ber tflatw emer~ntin grunbtegcn ran. ~ie tThlev in~taIiafunbbea~aIbennid~tvngefd)ic~t/ wiebeef~beun1ff m~om/ V~e4pcI8jDenebig/51ovmi~/(ConftantunopeL/zc.fl)dd)calit in~jtaiia contva~

fcI~ter/mnbred~tin grunbgelegt/ in groffer form getruc~e/vnb.mir3U 1~anben~ommenfhnb/wieid~lie bairnapct~inbi~

fl~erc~(~bcrg~rI~icun)georb~netL~ab.

~n 9en ~tz~t~inbct.

c~~ ~teu ober anbcre groffc 3’i~uten;Io i~bcr1merc~t,onCIfl(*!(~oIszmna11bcr anbirnf~c~cr~irutenifrUcn cpnBcbunben wirbenwie anbere tanbta$lcnibarumbfic auc~aUsvcgcnba5 mucctbco

~Quatcrnober Zritcrnrnnf)abcn, obcr fk~naUciniev~cDie ~41 vnnb 6sgnarnrau~s

wafcn.

Fig. 5. A pageof preliminaries,includingacknowledgements[8].

Page 8: Computer-aided production of scientific documents

170 J. Brandi / Computer-aidedproduction ofscientificdocwnents

2.2. Markings tions, etc., the numberinghas also to reflect to

The aspectof logical structurealso pertainsto hierarchy.the marking of individual entities of the docu- Such markings often are referencedin otherments. Usually, such markings are numeric; we partsof the document,be it apagenumberor athencall themarkinga “numbering”. Sometimes, sectionnumber or name, or be it a footnote athe markingsconsistof alphabeticcharacters,or literaturereferenceor a remarkthat is referencedof specialsymbols like ~ ~ ~, ~ morethanonce.The commonlyusedtechniqueisetc. Quiteoften the choicebetweendifferent types to associatethe automaticallygeneratedmarkingsof markingdependson the environment. to someuser-definedidentifiers that remaininter-

The valuesof suchmarkingsarenot absolute, nal to the input text. It is theresponsibilityof thebut relative; usually they are simply sequential, userto keeptrackof suchsymbols.A good systemstartingat a standardvalue,e.g. 0 or 1, or, as the will assistthe userin keepingtrackof his identi-casemay be, at “A” *2 “f” #3 Numeralsmight fiers (If~TEX providesits LabL St - t ex file forbe written in different representations,arabic or thispurpose).roman, both uppercase(1987, MCMLXXXVII)or lowercase(1987, mcmlxxxvij); (cf. fig. 6) char- 2.3. Bibliographiesactersmayappearasuppercase,lowercase,greek,etc. Bibliographiesobviously representa particular

Many systemsprovide automatic marking of problem due to the wide variation of quotationcertaintypesof individual entities(they also pro- formats. While a running text has its “natural”vide, of course,meansto overridethe automatic orderof words governedby grammarand syntax,device). This is a remarkableprogressover the a quotationconsistsof many entitiesthat canbeautomaticnumberingof pagesthat simple text- put into avariety of orders,andcan be typesetinprocessing systems offer. The numbering of different fonts. It maybehighlighted,emphasized,chaptersandsections,of tablesandfiguresusually variously aligned, abbreviated, etc. Authors’doesnotchangetoo often during the creationof a names,e.g.,may appearin a form like: [full-chris-document,so thatsomeof usemight not think too tian name]+ [family-name], or as: [family-name]highly of this progress.But we all know the large + [abbreviated-initials], etc. Multiple authorsmayamount of boring work and the high numberof be fully enumerated,or elsetheauthorlist maybeerrorsthatwe incur, whenajournal’s editor insists abbreviatedby the“et al.” token.Journalvolumes,thatbibliographicreferencesbenumberedsequen- issuenumbers,pagenumbersandpublicationyeartially in an article. During the writing of the may be arrangedin various ways. The title of amanuscriptswe had found it necessaryto add or papermay be arequireditem of the quotationorremovea referenceor to merelyshift one around not. The preliminaryproceedingsof this Summerandnow suddenly— if we do it manually — we school [9] show the greatvarietyof bibliographyfind ourselfin amessof numbersthat havebeen styles.replacedand/or other numbers that still await The reason why such a variety is possiblereplacing. without affecting thelegibility too seriouslyis that

Automatic generationof sequentialmarkings a literature quotationis a highly structuredpiecefor variouspiecesof materialis thereforeahighly of information. The merefact that we refer todesirable,if not indispensablefeatureof a docu- eachof its componentsby name— as I did abovement processingsystem. If the material to be — indicatesthis structure#4~marked is hierarchicallystructuredas with sec-

*2 Alphabetic charactersarealsoassignedto thesequenceof

logical entities in an ascendingorder, like “A”, “B”, ~ Most of us probably keep our collection of literature insomesort of adatabase.Look at theitemisationof sucha

“C~ Evenwith suchspecialsymbols,thetypesetters’tradeseems database,or at the “field names”of apublic bibliographic

to usesometraditionalsort of ordering, databasefor more corroborationof this aspect.

Page 9: Computer-aided production of scientific documents

J. Brandt/ Computer-aidedproductionof scientificdocuments 171

~t$~t~ attet~fitntemrnrn~4n~bcrni ~cpfci4ummcn/ ~nt~ctct~rni ~mfc~aff~cn/‘D6Ic~ctit,êcccrcn,~1cn/~D&vff~rniøcb13flcvn,CL6ftcrn,W&Lbcn4h*g#

~ attcr1~namm~afftigm~5cfcf~ic~uwfD rn bufcln ~flcrcrber ~ofmograpf~cpbcgrif~

fmtvcrbrn.

~acasas*a ~cccrtJ~I bft~cflenae~enb~~C(bniiburgb~aetau bcam.~4~ ~flthj

F

~ bu ~rnmb ~nnaanrna~a. c1D.s~ina. en ~ ~peUcn

~n£ateln .tq.uf~raee~ ~ ~~ienannt bccrfl ~mnnnic~ ~cf~ flbvn~/ vnb inli ~a~ ~pr4*~iunOnrg cciv

~‘ 2~a~~33~fpeungccc~ ~2~bcrt~aabee ~~lteca. ~~I~mncuerbtf4afft*it~vn’I(obo, ~Ippin~e~erXrtig

~ ~UD bII~Cbt~IUd~‘~‘ an. crobnt baUj ~Xppen4.derinerbnngeftbf.g.a keim.itun~crfentbnntau,b~, bccccir~si ~ftiifp~i~ bccl~ ~ n~n bea ~pnir ~Iccqc*

~Xzenmbria ccim~.~?Ucctt~ ~peU~nZn~pd$u~OctpW~u~M~tbrt~~ anfa.ggenom ~embnrn bn.~rnfl~nn.~i~ngen ~nn ~ptila6teUi~etb

men ~?cc~ ~ ~ j~ fnrnm~ 6.. ~pLs~.ffatt en~ aiff cth~~2thnoba the n~erg cccc tag bet

COn Obctri,)lotcmaje W1CUSj. ~2ccc. ~ ci.et~nt tccl~1 ~1VtcIenb(sf4xnnmelha~* Zcatf4ianbc~cccC 2(lgóta bCdil$ CCCCe~

2~thein~tt~sieWnfthafft bcccth,j ZI~m~~cnnnb B4dcit ZpIUa baa~in~gie*ib taLe en erfin.sben2tetilaIanabZarfu.La CLaim Zagnó 4~nt b~CtC%’%d nub ~.genomam ectrini

banapatogcbamm ~cCtttv~ ~Ogó~~~~4(b~tu bcc~tvfij ~Lq,iUda cciu•. ivirb arobert eccnfl2icbetlio e~a~ctIOtibnn~ .bccl~tel ~taLr Sticcirrtn lLrabLa ~ccint,g ban ftdnLg ~itcasi$

innun ,nab non inca ctfrntbnu ~1ta1rbie~iaet mit ftzmpt ,brerXbc.n- ban etnôbig / nubinarulabnil.~euañC

~?cc~rcuj ftaf1~iung ~cccrcu~ ~cecit. baa t~b42~bettao~,erbgriontmc cccciricu, a.berenattgr~ffe ~?CCttCL22(radn~a~ bini ~taubenX~4m ~?cci~nttj 2irbnn btcln’t~2ibdOn ~~tfa~ 6th. 2taebge.~5e~ervnnb nOt ~iaffcfbafft ~tcnb~.eeno~n~Iebnibnvgdafl~~cclnrn

ccctln, c~iLj. ~e~onetat~Cait2the1.~atb ben benZeucfthen cctcfri-.n, X5,bIa~ ~ ~n.~ijtt~i,g4af~m 3actatmic an ban ~6mif~ ~n4 tommC~bnndn~niner6$iate ~tcccn ccctl~tij c*U~.iompian~randietn~ ibib.

no tngcfuabmtunrben ~ ~iaenab tLimpffnt tao ~ef~ctgnb,a~Linca ~(rdbIIII tattuutiinan,uebbn.3~bmifthenOtci4,3cic~tn ben ~accct~,tej Xrgt3Ba$~er bccrcirCCC~P ~cipcafpa4 bcctit Xxg,ppcu na.fir~6i~ea

~Xbict.Item nab~9?igtrnm tCCti 2t1p(jonfu.2(fir.nnmn. I5~i~ZtIa ~ttcmjit ~boip~,.non ~aff4n tnirb eefc~tagcn ltenbustg Cu. ~ bccccilt; Zrtaba..tanbt ~cccitnti

bCtc~ 3Otniu( bccccen 3fruiimim2(bflf4 ~?eer (totfi ~Utripp bc(st~ ~rL~ntdLa~ttbnnØ2tiien nn btiij Xmteiber cccc;cnuj 30U0bit2tictaium ~cjCiLiL.ttu$~1ffcnlate~iegcfangenn,erben ~?CCcIi. ~iiifleeienenau ~ ~rieu.berg bccetn~

t~tccccn.j ?ui~eimevesta ~~cn Xtlcn bCCIti*ma.on in ~i4bcgreiffi ~Ii.rUi 7AinaioluenØteLftbau’2Denber ~O~ccanr ~nnenta banUcla ~?cc’noitt

2ifrica ban Ucla ~CCCX~Tti)3tmbnrg betecactaa(~re~bconcxafactur ~Lran.iaban gacife~tfrlca malt Iota..bafunbetn L&nbean,Z~.c bcc#ratti ~(racn4ede.XUeg be~betenact

tea nnnb tnunberbaru4enbinge. ~mbergcrZretannbetautbi~afftigtettgc bili~CCCrViti genl~umivagca.~cr2eabcccttxtlt 3lcmnufcctenbcge4ttettfhttt

3(frica b~nnet. ~ccccriiij ~jmcru~ ~?CcCt~vi Liii~Lgachr~ manfbr~6ttftr ~cc~Xttt 3imienn cuteeau ct~ Zrnburg CCCC(n2(gaunn,n c(cctrrn’it, ~immoninZcapct tie bent tiwctcn tiL’~a ~traftetni gelnunum bcccrcif?(gIarbee~i1th6ffnnb ~atriat~cie tautpe ~?CCCXC 2tnCro$agtma.fbr ‘l3lttter ~

reglertI,iaben Cct’t 5mmoenfort cccemv ~Lrnu.QBafrcc titer2~grlgenCumci.ec.tc Ct~Itt~ ~jmo;ba~ ~ ~ bcccrci;j ~roia3(baulon bccc~iij ~ ~ ~‘4nagou*witZac1~~ionieni Qattbaieteicni2~baufennon taintge(hffeat bcccc~eem~~ja(ferba~ bteecaualt Inapt-bet ~e. ~a4onuIen/maolAr etett nub Boa..

.t1at~~2J3ai1ar bcttctr f~eeebnng cItiz.vj Ievnbc~buient,abcnCCcctO bccctv;i~ 2(mueatesbet~. ZOrthfib .Ienfea ~ccmij !Xnngente be.itbaugneub.an(au~inin.e

man bcutliti tap f~.ttin Lanbc. ~ccgtjj en bib in t~nfcrnrjut T{4iUt l*.ibilltjliblth.nfft, t’nnb mvarumnb all, gcnannt 3~muratenbet i.Zi%rd~f4£ ~ct(u Zreagonecr etobcrn ban 3~ct*eitaiutn

2Luth,na CCCaj CCLII1~?ccIrn~ntj 2tnbetnath btcti* ~rta. bit enactmit Inapt~tet Cefubiet.

3tlbcatun ~Ø?egann btcc~it.bccclrir. ~(nbtaa bciiuj bang curb�~nbeec~niie~itbuu~ CCXIIL ~tii~~~t ~CCCCIj

2hb.on ~Iltl~ ~ ‘~cmrtib bIwpiabc.~l~Oia~obcr~~fl4ob.tb..Znig clitlit ~Oinoitbiugn ~Iubotp(tLt~ematai05mb X~taban tmeumt

/ 05i. ~iabolp~. non Sabfpurq en Oafei CCCCr~t,~ ~fmebe. aubcrnlfcl(.bet~cmc%.i~nieo~mnmirbabmircbcritlniq,anntrb ~(naoubetannetbiog bCCCIt bang ~kcIrCLrrtbbt nolenten 23iubcrn ecu. br.ttj 3Anln beeBug m4er it atfo geuluiet meeb ~ti,na ~?ccct1t#.~umuib

2itbttcf,t, .St6iugn in 2ingcrn ~iftO*9 cctnej ~srnnca titern~crc, 2tnti.ibia ~?cccltLb ~ffnrter am ~in~ ~uba nbd gelutien

$3 2tlbiec9i, Q.ofltrif4enng ~?cniij ~tutnuffbleecaic, fompc berfclbtn !~c, ~3?cCc~pin~bcc~v;lj l4sesbnitg bccrrv,.bcctrn.l ~el~on

):( 3Ac~aecno

Fig. 6. An indexpage;notethepagenumberin lowercaseromannumerals[81.

Page 10: Computer-aided production of scientific documents

172 J. Brandt / Computer-aidedproduction ofscientificdocuments

Functionsfor structuringtheseentitieslogically recursive.Weoftensay that wereada book “diag-and for defining different stylesof renderingthem onally” at first ~, andlaterread“morein depth”.in different environmentsare indispensablefor A logical consequenceof this approachis thesystems that pretend to effectively handle this presentationof a documentwhich initially onlykind of problems. displays the general ideas, and which later, at

Bibliographiesposea further problem due to strategicpoints and at the reader’srequest,givestheir natureas “ancillary materials”: the problem more detailsto that point ~.

of placement,as is discussedin section2.4. When the substrateof the documentis not astatic two-dimensionalsheet,but a more flexible

2.4. “Ancillary material”: footnotes,remarks, refer- device, this leads to the idea of “display-on-de-ences mand” which is the essenceof Hypertext(andalso

of similar systemsthat sometimeswerecreatedasRemarksare the spiceof a text, andsometexts someextended functionality in on-line retrieval

are certainly “overspiced” in this respect.On the systems).Hypertext,as yet, is not too widespread.other hand, they are an indispensablemeansof It is hard to predictwhetherthe systemsthat arekeepingthe mainstreamof texts free from sub- currently on the market, will really achievecorn-ordinate, albeit indispensable,material (see also mon acceptance.The problemsthat theyposeonsection 2.5). If such ancifiary or parenthetical documentinput, seemrathersevere.materialwerephysically includedin thetext, read-ability would suffer severely#5 Technically itmakeshardlya differencewhetherthis materialis 3. The man—machineinterfaceof explanatorynatureor whetherit actuallymod-ifies the text (as is often thecasewith the critical On the practical side we are, at present,wit-remarksin linguistics). Practically, thereis a dif- nessing a controversybetween interactive andferenceaccordingto the physicalplacementof the batch systems.Such a classificationis not reallymaterial: if it is appendedto the end of the fundamental,sincebothcategoriesof systemsservecurrentpagethenthereis the problemof availabil- the purposeof producing close-to-perfectresultsity of spacethat hasto besolved,but theamount with minimum of effort. It might be a subjectofof floating material remainsrelatively moderate debatewhetherthereis — or hasto be — a closebecauseit will be dumpedat the earliestopportun- correlationbetweenthe functionality of a system,ity. If such material is placedat the end — as is and the degree at which it can “immediately”mostly the case with bibliographies,but some- presentthe effects that result from a useraction,timesalso with otherparentheticalmaterial — then and also the degreeat which this presentationisthe amountof floating materialmaybecomequite more or less easyto interpretby the user.large. Obviously, the more such a user action has

In both cases,an associatedproblem of cross- far-reachingeffects, the moredifficult to programreferencingoccurs,becauserather frequently, the andthe moredemandingof resourcesWYSIWYGsame remarkis referencedmore than once;with will becomes.The situationis somewhatcompara-bibliographicreferencesthis is almostthe rule. ble to programminglanguagesand their transla-

tion/ interpretation. The most interactive one2.5. “Recursivetext”: more or less detailedpercep-tion ofdocuments

*6 this, I assume,isjust anotherword for theattemptto graspIt it is true that learning is a recursiveprocess, somegeneralideas(the Germanexpression“llberfliegen”

then readinga nontrivial documentalso mustbe is quite appropriate).*7 We notethat with the traditionalsubstrates,like paperetc.,

*5 A TEX input file is a good illustration for this, asfootnotes this kind of structuring was achievedby typographical

indeedhave to be insertedin therunning inputtext attheir means,but alsoby useof parentheses,footnotes,remarks,point of reference. appendices,etc.,cf. section2.4.

Page 11: Computer-aided production of scientific documents

J. Brandt / Computer-aidedproduction ofscientificdocuments 173

amongearly programminglanguages,BASIC, re- they force him to do so. As a consequence,userslied on interpretersand sacrificed the aspectof of batchsystemstend to specifythe logical aspect,global consistencychecks. i.e. they will not attachany typesettinginstruc-

On the other hand, we now have some nice tions to a logical entity, e.g. a title, since theseexamplesfor a dramaticincreasein user interac- dependon the environment; they will insteadtion evenwith compiling languages.Let me men- eannark the text of that title by a classifyingtion BORLAND’s TURBO-PASCAL where we token, which says “this is the title “. The samefind a nice implementationof the basic idea of principle holds for many other kinds of logicalstrongly interlinking the program source editor, entitiesof thedocument,suchas tables,footnotes,the compiler, and the runtime-system,concurrent figures, etc. Texts of abstracts,remarks,acknowl-with the techniqueof keepingall this, asmuchas edgmentsandliteraturereferencesare particularlypossible,togetherin main memory. conspicuousexamplesin that they may be typeset

Consequently,the boundariesare notrigid nor as footnotes,or be collectedin a separatesectionwell defined. Any classificationis thus basedon of the document,or evenoutsidein a separatethedegreeof compromisebetweenfinal qualityvs. section of the embracingenvironment.Figure 5speedandeaseof use. It alsodependsveryheavily (lower part) shows an example of a technicalon the technical progress, and-somewhat— on remark to thebookbinder,regardingtheplacementfashion! of doublepageartwork (the problem of including

graphicshasa long traditionindeed!).3.1. Batch systems

Batchsystemspretendto aim at a maximumof 3.1.3. Cross-referencingquality and functionality, and at a high degreeof Mostdocumentscontainsomeamountof inter-portability at the expenseof user interaction.To nal cross-referencing.Suchcross-referencesmightsay the least, they attempt to provide the full be independentof the context, but more oftenrangeof tools to achieveperfection.The quality they to dependon it. We might referencea certainthat is actuallyproduced,dependsto some degree equationby its number:this numberwill changeifon the effort that theuseris readyto invest.Effort we decide to treat another equation first. Weis required both in getting acquaintedwith the might refer to acertainchapterin a bookthat we(high degreeof) functionality of the system#8, arewriting, but again— as we all know — we quiteandin inputting the commandsandproviding the frequently insert chapters or change their Se-computerresourcesfor their execution~. quence.The sameholds true for literaturerefer-

ence, footnotes, theorems, figures and tables.3.1.1. User-friendliness Cross-referencesto pagenumbers,obviously, are

Oneaspectof user-friendlinessof suchsystems extremely context-dependent.Thereforethey areis theinitial (minimum) amountof systemknowl- usually not eventypesetin galley-proofs.(As toedge that the user has to master, before he can this text, its page-numberingobviously dependsproduce a document at all, albeit of moderate on the page-numberingof the surroundingen-quality. Another aspectis the easeof increasing vironment, i.e. the other articlesin this journal.)his sophisticationgradually,i.e. how easilyhe can Batch systems provide facilities to referenceadvancesmoothly — and not by big boundsand suchitems by variablesrather thanby constants,leaps— to the full useof the system’stools. and to assignvaluesto thesevariables in a con-

text-dependentfashion.3.1.2. Supportof logical structureof documents

Batch systemsallow the user to specify the3.2. WYSIWYGsystems

logical structureof his documentand sometimes

#8 This might becomparedwith thepurchasepriceof a piece WYSIWYG systemsstress interactive editing

of equipment,thuswith an initial investment..., throughimmediatevisualisationof the final docu-*9 . and this would correspondto operatingcosts. ment. Underthepresentconstraintsof affordable

Page 12: Computer-aided production of scientific documents

174 J. Brandt / Computer-aidedproductionofscientificdocuments

computerresources,WYSIWYG systemshaveto processingsystem.If they are offeredat all, theycompromiseheavily in favour of executionspeed, requireanadditionalprogramor programmodulei.e. responsetime, therebysacrificing quality and to be run in batchmode.flexibility to someextent.Multi-window multitask It remains to be investigatedwhetherexistingsystems may give the user an impression of highly functional batch systemscan be modifiedWYSIWYG performancealthough they actually easilyenoughto executesomeimmediatelyvisiblerun as a clusterof separatebatch-typeprocesses; processingwithin a limited — possiblyuser-defina-insofar they bridge the gap between the batch ble — context.approachandWYSIWYG systems.

4. The man—paperinterface3.2.1.Portability

WYSIWYG systemstend to sacrificeportabil- At the beginning I mentionedthe man—paperity, becausein the interestof short executiontime interface. Some time ago “the paperlessoffice”theyrely very often ratherheavilyon machine-de- was a popular slogan. We all know that todaypendentcode. They might be portablewithin the thereis more paperaround in offices than everscopeof a certainoperatingsystemor acrossthe before, in spite of the fact that thereare morerangeof a certain family of hardware,but quite CRT-displaysavailablethanever before.often they evenbypassa commonoperatingsys- It is worthwhile to at least cite some of thetern’s commandlayer and interact directly with claims that are madein favour of paper,or againstlow level hardwarefeatures,e.g. manipulatekey- it. All of thesemaybehighly debatable,extremelyboard interrupts, CRT-buffers, disc controllers, subjective and variable under technical progress,printer driver neverthelesssuch aspectsstill play an important

Sucha high degreeof hardwaredependencehas role in the scientists’ strive to producedocumentsbecome acceptableto some degree due to the rather on paper than on anything else, and toadvent of highly compatible, widely available producethemin thehighestqualitypossible.computer families, e.g. DEC’s PDP-11, or VAXmodels,and throughthe appearanceof de-facto- 4.1. Psychology: the perennialcharm of the scrip-standards,of which the IBM-PC-XT-AT family toriutfl.(andclonesthereof)is agood example. Paperseemsto havea particular flavour to it,

perhapsevena fascination.Wemay askourselves3.2.2. Context-dependentoperations;cross-referenc- for someof the reasonsfor this. Certainly,thereising thepsychologicaleffectthatmedievalmonksmust

WYSIWYG systemscompromisevery heavily haveexperienced~ in their scriptoria whentheyon the extent of context dependenceof the ap- hadfinishedabeautifully looking page,or a corn-pearanceof the piece of the documentthat the pletebook, firmly and durably bound in leather,user is actually interacting with. It is not too destinedto last for ever ~‘. Paperis visible anduncommon that a user action only affects the (relatively) durable! In moresoberterms,we thusappearanceof one single line of text, and that haveto consider:reformatting one paragraphalready requires an • greaterdurability of paper #12 vs. volatility ofexplicit directiveto be given by the user.Optimis- magneticstoragemedia;ation of the layout of the current page,or of alogically connectedsequenceof pages,such as a #

1OA feeling that probably they shared with Gutenberg’s

chapteror a completedocument,is usuallyrather workmatesand earlysuccessorsin their officinen.underdeveloped. *11 , . .or very long, let us say: many generationsthencefor-

ward.Document-wide operations, such as facilities ~ ...although average20th-centurypapers are much less

for logical cross-referencing,or automaticproduc- durable than olderpapersdue to higheracid contents;in

tion of indexesor of tablesof contentsare rather fact, largelibraries todayarein dangerof decayingwithin aseldom part of the WYSIWYG-type text few decades.

Page 13: Computer-aided production of scientific documents

J. Brandt / Computer-aidedproductionof scientificdocuments 175

• higher decorativevalue of many paper docu- Certainly, the limited capabilitiesof computerments (“ more prestigiouslook” of a bound printing equipmentat that timedid notverymuchbook) vs. “crude technical look” — or even to encourageresearchin methodsfor which noinvisibility — of computerstoragemedia. affordablehardwarewasin sight.So, it might not

be a merecoincidencethat researchtook an up-4.2. Practical aspects:how real are they? swing with the advent of ink-jet, laserand 24-

needleprinters.They at lastpromisedfreedominBut there are also practical considerations, font, size, and orientation, and a practically un-

Without any claim for thoroughness,let me pro- limited arsenalof specialcharactersand symbols.posesomeaspects,such as: Obviously, the mostpressingneed— also eco-• immediateavailability of the documentvs. re- nomically — existed in the publishing of science,

sponsetime of computer; i.e. physics, mathematics,linguistics, chemistry,• parallel presentationof severaldocumentsvs. etc. In theseareasof science,thereseemedto exist

limited (or unavailable)windowing facilities on somethinglike a set of typographicalstandards,orcomputerdisplays; at least of commonlyacceptedusage.This, how-

• higher information density(and graphicsqual- ever,wasonly partly true, sinceusagevariedwithity) vs. limited size and resolution of CRT time and cultural environment,as illustrated byscreens; the comparisonof typesettingstyles in a mathe-

• better readability through higher contrastand matical journal over different periods in ref. [7,throughgreatervariety of typographicalmeans pp. 3 and 5]. The creatorsof SCI-TEX, SCRIBE,for differentiating between different compo- and finally TEX — which certainly is a majornentsof the text, or document,vs. low contrast milestonein this field of research— presumablyand uniform appearanceon CRT displays had some difficulty in formalizing such de-facto-

standardswhich were quite complex, and per-4.3. The art of typography: expert knowledge indocwnentproducingsystems mitted agreatwidth of variationdueto historicaldevelopment,traditionsof local typesetters’habits,

Although local, and sometimesrather specific, and personal gusto of editors and authors.Thetechniquesweredevelopedfor thesolutionof some terms “iesthetically pleasing”,or “easy to read”of the problemsmentionedabove,therewas — in wereno absolutevalueswhethertheywereappliedmy opinion — not too much of systematicand to font designor to documentlayout.truly scientific researchinto the fundamentalsofcomputeraided typographyuntil about one de-cadeago. 5. Presentstatus

The reasonsfor this are not veryclear,becauseademandhadexistedlong before.Perhapsit wasthe double pressureon the individual computer At presentwe are witnessinga growing accep-scientist who was preoccupiedwriting a docu- tanceof WYSIWYG-typeof “desktoppublishing”ment,andat the sametimerealizedthat heneeded programs.The initial enthusiasmof a noviceuseran up-to-datetool for its production,that led to about the virtually unlimited degreeof freedomthe fact that the tool received not more than for embellishinghispaperworktendsto fadeawaysecondaryattention.Perhapsit was also the type- whenhediscoversthat thepossessesa tool buthassetters’guild who tendedto regardtheir tradeas not learnedthe tradeto which this tool belongs.an art — which it certainlyis to someextent,but He then looksbackto computerscientists,askingnot exclusively — and who were not very much for assistancein professionaltypography.We in-inclined to communicatewith an outsider*13 deedhavetools available to assisthim with the

productionof a professionallydesignedand con-#lSwhOseprofession,moreoverwas threateningto eliminate sistently produceddocumentof non-trivial size;

somejobs from his trade. TEX, togetherwith macro packageslike LuTEX or

Page 14: Computer-aided production of scientific documents

176 J. Brandt / Computer-aidedproductionofscientificdocuments

AMS-TEX, may be considereda leader in this [21Le Figaro(15 December1837).

field. [3] Johann Jakob Christoffel von Grimnielshausen,DerAbentheurlicheSimplicissimusTeutsch (JohannFillion,There also existsnow a machineindependent Monpelgart(= Montbéliard), 1969).

representationof typographicalinformation: the [4] Michael Huck, Typeset(University of Minnesota(1972).

DV I-file for transportationof document-layout, [5] Heijo Klein, Sachworterbuchder Drucktechnik und

and the METAFONT language for font design. graphischenKunst, 5th ed. (DuMont Buchverlag,Kdln,

This, too, canbeclaimedto beamajor advantage 1981).[6] Donald E. Knuth, Bull. (New Ser.) Am. Math. Soc. 1

of the TEX system.(As a vehiclefor the publica- (1979) 337—372.

tion of computersciencedocuments,themerits of [7] Donald E. Knuth, TEX and METAFONT - New Direc-

WEB andTANGLE still await full exploitation.) tions in Typesetting (Digital Press and AmericanWhetheronewill actuallyuse someTEX-based MathematicalSociety,Bedford, MA, 1979).

systemlike LkTEX, or some similar program is [8] SebastianMunster, Cosmographey.Oder beschreibungAller LanderHerrschafftenund furnehinestenStettendes

not essential.In my opinion, one greatmerit of gantzen Erdbodens samptjhren Gelegenheiten/Eygen-

TEX is that it has offered us a language for schafften/Religion/Gebreuchen/Geschichtenvnnd

describingcomputer-aidedtypography, and be- Handthierungen/etc.(Basel, 1588).

yond: for discussingcomputer-aidedproduction [9] JaroslavNadrchalet al., 8. EuropeanSummerSchool on

of documents. Computertechniquesin Physics,PreliminaryProceedings,Skalsk9Dvür (CS), Fizikhlni üstav,(~SAV,Praha(1989).

[101 Barbara Salberg-Steinhardt,Die Schrift: Geschichte—

Gestaltung— Anwendung (DuMont Buchverlag, KöIn,References 1983).

[1] “Portr’át des Genius: Honoré de Baizac” (Marion vonSchröderVerlag,Hamburg,).