PrePress Lout and Multilingual publications

[Ian Carr-de Avelon]

Dear All,
I have mentioned previously these 2 things which interest me, but after 
a fruitless web search for a good available alternative and a look through 
the source code I have decided to come back to see whether my initial
ideas make sense to people who know a lot more than I do. There are a
lot of ways of doing what I want, including starting to write a whole
new package, I'm mainly interested in a way which will reduce the need 
for my hacking time, although I do place some slight importance on 
computer efficiency and elegance.

The first thing I want to be able to do is produce PostScript for producing
film for offset lithograph printing of a booklet. I know many of you 
understand better than I do what that involves, but I'll put an explanation
here for those who don't.

At an offset printing works each page is passed through the press a number
of times, each time adding one colour. The booklet I have in front of me
has a full colour cover, so 3 primary colours and black are each printed on
it. Other pages have 2 colours (termed spot colours), so for example:
one plate has all the text except the first letter of each paragraph,
another has only the first letter of each paragraph. By printing the
first plate black and the second plate red the whole text is printed,
with just the first letter of each paragraph in red (anyone got a lout
macro for this?). The plates are delivered to the printer on a large
plastic film which has several pages of the booklet arranged on
each sheet of film. As well as what must be printed in the book there
are also lines to show where the pages should be cut, and small squares
at various levels of grey for the printer to  setup the press correctly.

It now strikes me that I could get this quite simply. If I add an extra
command line option to lout like "lout -Pc" to mean that I want the 
cyan plate, then a hacked lout postscript backend could, instead of
putting a postscript colour, calculate a grey level and put that. For
spot colours in the general case the backend would need to know the 
colour of each of the plates to be used, so it could include a letter
if this plate's colour is the closest (in practice a direct match), or
omit it. For this case I could get away with black going on to one plate 
and any other colour on the other plate. 

Running lout like this 6 times would give me 6 postscript files with all 
the pages, suitable for passing to an A4 laser printer. With PSutils I
could select the individual pages from these files and place them in
a new postscript file to print several pages on the large sheets of 
film. Obviously if I could persuade lout to put each page in a separate
eps file during a run of the booklet, I could use lout to stitch them
together and produce all the cutting marks etc. which would be a lot
more elegant. 

Thinking about complications, it strikes me that I'll need to hack loading
of included illustrations. If I have photo1.eps included in the document,
once I have finished proof reading and give the options for the colour
separations. Will have to include photo1.c.eps on the cyan plate etc.
The GIMP manual explains how photo separations can be made. I'm not
entirely sure what the effect of outputting lettering with a grey level
of white will have. Eg, if I have a gray (ie real grey on the finished 
page) box with  red text on. Using the hack as I have described, I
would get the grey box put on the black plate, but seeing that the 
red text didn't match, lout would output it as white. Will that remain
a grey square on the black plate film, or will it end up grey with 
white writing inscribed in it? I'm not sure whether the white letter 
spaces left in the grey box, for the red letters may not be what is needed
anyway.

The second thing I'm interested in is publications which are produced in 
a number of languages, where the intention is that the translated content
takes up more or less the same place in each version, although the amount
of text in a translation may differ quite substantially and an extra 
footnote or two may be needed in one translation or another. Looking
at how lout works, it seems you would need to produce a galley with
the text for a small section in each language, assign the space needed 
for the largest text and then use that space as well as it can be used for 
the smaller texts too. Would it be necessary to hack the source code
to achieve this, or should it be possible to produce macros?

Yours
Ian