[net.text] Navy Document Interchange Format

ted@imsvax.UUCP (Ted Holden) (10/09/85)

My apologies to anyone who has seen this article more than once.  We've
been having major problems with usenet in the D.C. area.  Since Sept. 15,
when the CVL computer at U. Md. was taken down with no warning to anyone
below it on usenet (which is most of the D.C. area), very little has
gotten through one way or the other.  For awhile, it will be difficult to
tell what has gotten through and what hasn't.




               This  article  concerns word-processing document transmission
          and  also  a  typesetting  and  graphics  system  which  is vastly
          superior  to  TROFF  and  other  such  archaic packages which are,
          nonetheless, still in large-scale use on UNIX systems.

               The need for inter-vendor transmission of DOCUMENTS in such a
          manner  as  to  preserve  word  processing  FUNCTIONS,  as well as
          page-image, is fairly recent.  The Navy saw three  or four reasons
          for wanting  to have  the capability:   1. as a  guarantee that an
          organization doesn't lose a ten  year  base  of  documents  if one
          particular  vendor  should  go belly-up (i.e. they could translate
          their store of documents  to some  other vendor's  format), 2. for
          transmitting documents  around the  office or  around the world as
          opposed to printing 100 copies of a document  and MAILING  them to
          100 people,  some of  whom could be in Germany or New Zealand, and
          3. as a means of mixing  and  matching  the  talents  of differing
          kinds  of  office  equipment.   It  is this last possibility which
          should most interest Unix gurus, as I shall  describe momentarily.

               Pairwise  routines for word   processing   document    format
	  conversion, whether  in  the  form  of  contractors  or  of "black
          boxes", are   expensive,  a pain to  use,  and      generally only
          support a few popular  formats.  There  is a  further problem with
          DCA/DISOSS,  which  is  IBM's  solution to the document conversion
          problem;    DCA  itself   corresponds   almost   exactly   to  the
          functionality of  a 1965 typewriter.  It is essentially impossible
          to write reasonably accurate translation programs  between DCA and
          the file  structure of any reasonable 1985 word processing system.
          Documents  being  transmitted  from  system  A  to  system  B  via
          DCA/DISOSS can be counted on to arrive in mangled condition.  

               A user  can only  feel good about sending a document over the
          lines if he can count on it looking right when it gets there.  You
          particularly don't  want tabs,  decimal tabs,  or indenting to get
          mis-translated or end up on the wrong columns;  that  could change
          the meaning  of a  document.  This  years numbers  could end up in
          last years column,  profits  could  end  up  looking  like losses,
          debits as credits.....

               The  United  States  Navy's Document Interchange Format (DIF,
          not  to  be  confused  with  Visicorp's  DIF,  which  means  "Data
          Interchange  Format")  is  the  best  available  solution to these
          problems.  The Navy DIF was deliberately  set up  to correspond to
          the  file  structure  of  a  reasonable  1985 word processor, with
          functionality  chosen  as  a   maximum   common   subset   of  the
          capabilities of  the products  of the Navy's largest vendors.  The
          DIF's functionality was also chosen so as to be easily "reachable"
          by translation  programs, so that DIF translations between diverse
          systems tend to be surprisingly accurate, good enough, in fact, to
          send documents  over the  wire and  expect them to look right when
          they get there.

               The most interesting possibility which all this has  to offer
          to  UNIX  types,  involves  the  Xerox 8010 "Star".  This, or it's
          newest incarnation, the Xerox 6085, soon  to be  announced, is the
          world's  ultimate  typesetting,  graphics, and publication quality
          word-processing system.  Anything which  can  be  done  slowly and
          painfully with  TROFF or any of it's derivatives, can be done with
          a few  keystrokes  on  the  8010,  exactly  like  using  any other
          word-processor functions.  The 8010 beats the hell out of TROFF or
          any  other  such  package.   The  8010  incorporates  a  1024x1024
          graphics  screen;    typesetting  functions  appear WYSIWYG on the
          screen, and  its  graphics  are  good  enough  for  CAD-CAM.  Math
          symbols,  Greek  letters,  the Cyrillic alphabet, Kanji characters
          etc. all come easily to  the 8010.   Anyone who  has used  an 8010
          would get violently ill if he ever had to use TROFF again.

               Now,  the   8010  is  not  the  fastest  or  least  expensive
          word-processor in the world, and it used to be the  odd-man-out in
          Xerox's line-up.   Potential customers  would be  amazed by it but
          noted that it was compatible with absolutely nothing,  and figured
          they either  needed 100  of them  (which they couldn't afford), or
          they didn't need any, and walked  out shaking  their heads.  Given
          the  Navy's  DIF,  however,  since  the  8010 has DIF routines, an
          amazing possibility presents itself.  If 95 percent of the work on
          a given document could be done on a PC or some cheap UNIX machine,
          and the  document  then  shifted  body-and-soul  to  the  8010 for
          finishing  touches  (typesetting  and/or graphic boilerplate), and
          then printed on the 8010's laser  printer, then  the power  of the
          8010  has  been  magnified  20  to  1;    the  input load has been
          effectively removed from it.  An organization could thus make very
          good  use  of  one  or two  8010s, which are not that difficult to
          afford, especially since their price has come down drastically  in
	  the last year or two.

               Integrated MicroComputer Systems  Inc. of  Rockville, Md. has
          DIF  translation   programs  available  for  the  SSI  WordPerfect
          package for PC's, as well as for the  Fortune:Word package  on the
          Fortune  32:16  micro.   Either  can  be  used with an 8010 in the
          manner  just  described.   Other  vendors  whose  word  processing
          packages  have  DIF  routines  which have proven acceptable to the
          Navy include DataPoint  (VistaWord),  Hewlett-Packard  (3000), DEC
          (Wps), IBM  (Display-Write on  the PC),  AT&T (CrystalWriter), and
          CPT.  A number  of  other  firms  are  presently  at  work  on DIF
          software.  Anyone  interested in this software may call IMS at 301
          984-8343 and ask for Ted Holden, Dick Jeffries, or Gary  Evans.  I
          will also be glad to answer inquiries addressed through UNIX mail.

michaelm@bcsaic.UUCP (michael b maxwell) (10/12/85)

In article <423@imsvax.UUCP> ted@imsvax.UUCP (Ted Holden) writes:
>          ...the  Xerox 8010 "Star".  This, or it's
>          newest incarnation, the Xerox 6085, soon  to be  announced, is the
>          world's  ultimate  typesetting,  graphics, and publication quality
>          word-processing system.  Anything which  can  be  done  slowly and
>          painfully with  TROFF or any of it's derivatives, can be done with
>          a few  keystrokes  on  the  8010,  exactly  like  using  any other
>          word-processor functions.  The 8010 beats the hell out of TROFF or
>          any  other  such  package.   The  8010  incorporates  a  1024x1024
>          graphics  screen;    typesetting  functions  appear WYSIWYG on the
>          screen, and  its  graphics  are  good  enough  for  CAD-CAM.  Math
>          symbols,  Greek  letters,  the Cyrillic alphabet, Kanji characters
>          etc. all come easily to  the 8010.   Anyone who  has used  an 8010
>          would get violently ill if he ever had to use TROFF again.
Well, I have a slightly different opinion.  I've used NROFF (not,
admittedly, TROFF), and Stars.  The Stars are fine if you don't have any
fancy formatting to do.  BUT--if you want any of the following:
	1. footnotes (as opposed to endnotes)
	2. automatically numbered foot- or endnotes
	3. automatic section numbering
	4. automatic figure numbering
	5. cross-references to any of the numbers in (2-4) which will
	   automatically change when the original number changes (e.g.
	   if you insert a new figure between fig. 2 and fig. 3, and
	   later in the paper you refer to "figure 3"--now "figure 4")
	6. reverse-indent paragraphs (I'm not talking about paragraphs
	   like these, with a hanging tag, but rather about paragraphs
	   in which the first line is "undented" a fixed number of
	   spaces regardless of the length of the first word--e.g. as
	   in some common bibliography formats)
All of the above are common in "scholarly" papers, and I suspect in many
other disciplines too (TV scripts, plays, ...).  If I'm doing funny
alphabets, then yes, I think I like the Star.  But it's a horror if I
have some particular format I have to follow.  I suspect there are
other thinks it can't do, but maybe there is a work-around, so I won't
show my ignorance.  A Star probably could be made to do most of the above
by suitable reprogramming, although I've long wondered how a WYSIWYG WP could 
do #5.  Until it/ they do, I guess I'll have to stick with N/Troff for some of 
my papers.  I.e. use a screw-driver for screws, and a chisel for...
-- 
Mike Maxwell
Boeing Artificial Intelligence Center
	..uw-beaver!{uw-june,ssc-vax}!bcsaic!michaelm

guy@sun.uucp (Guy Harris) (10/12/85)

>                The most interesting possibility which all this has  to offer
>           to  UNIX  types,  involves  the  Xerox 8010 "Star".

An even more interesting possibility for UNIX types is the Interleaf
document preparation system.  It's more interesting because it runs under
UNIX on Suns; I believe it also runs on Apollos, although I don't know what
OS it runs under, and it may run on other UNIX workstations with large
bit-mapped displays.  It's a *very* fast WYSIWYG editor (it reformats
paragraphs *and* pages as you type), and supports graphics as well.

The Xerox D-machines run Pilot rather than UNIX, although I've heard
rumblings about C compilers (although ask anybody who's done a C compiler
for a vanilla Perq about how easy it is to do C compilers and libraries for
word-addressed stack machines).  Xerox has finally released the Xerox
Development Environment, at least for the 6085.

I don't know if there are DIF functions for Interleaf, but it's worth asking.

Interleaf is in Cambridge; I don't have any more information on them.

	Guy Harris

chris@umcp-cs.UUCP (Chris Torek) (10/12/85)

I beg to differ:  Star is a mess.  ViewPoint, however, looks very
nice.  (ViewPoint is the new version of Star, with applications
for it being written in XDE 4.0---a.k.a. XDE 12.0E---and offers,
among other things, vastly improved speed and many moveable windows.
Star allows only four windows at a time, and has many bugs.  Whether
ViewPoint has fixed the worst of these, only time will tell, since
it is not yet available.)
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@mimsy.umd.edu