Open Problems

Joachim Schrod
Sun, 17 Sep 1995 20:59:13 +0200 (MESZ)

At EuroTeX I made an error -- I agreed to collect the open problems
that have been mentioned on this list since we distributed the draft.
IMO, many of the stuff are not really problems, it's just that
somebody must sit down and must write it up. ;-)

I'm loosing my job in two weeks (and have no new one yet), so don't
expect much engagement from me in formulation issues.


---------------- included: open-problems.txt:
 -- One must not use texmf/tex// as search path, due to duplicate
    files. TEXINPUTS must be set for each format anew. (There may be
    reasonable defaults, but eventually that's the bottom line.)
    [Paul Vojta]

 -- PK files not generated by MF go in
    [Joachim & Karl]

 -- Where are binaries placed?
     o  Non-Unix folks don't like bin/<platform>/.
     o  Distribution of binaries over all kind of directories is not
	sensible. Placing binaries in TeX distribution dirs (as
	proposed by Christian Spieler) covers only a small part of the
	problem, that of the distributor -- where shall a TeX admin
	place the binaries [s]he has made locally?

	IMO, we should discard that issue and mention explicitely that
	we could not reach a consensus on that topic. We should
	mention that one needs platform-specific directories very
    [Christian, Pierre, Karl, sebastian, Joachim]

 -- It must be made more explicit that format, base, & pool files are
    placed in the <TeX implementation> tree. It's in the draft, but
    this is the most FAQ, the wording must therefore be improved.

	 Perhaps one might also need to add a rationale (a) why it's
	 there and (b) why we don't specify _where_ it is placed
	 there. [(a) since these files are inherently implementation-
	 dependend; (b) since we cannot enforce it on developers, we
	 would structure _their_ area.] We should made explicit that
	 we encourage developers/distributors to tell us where their
	 FMT files will be placed, and should add that information in
	 an appendix.

    [tons of people]

 -- An appendix should be added that outlines known structures of <TeX
    implementation> trees. It must be made explicit that this
    structure is for a specific version, is not fixed, and may change
    in the next version; it's just for informational purposes (to see
    what `others' have done).

	 Most questions concerned the implementation-dependent area.
	 People want to know something about this.

    [Christian, Pierre, Karl, Thomas]

 -- TDS Registry -- what do we do with it?

	I have to admit I didn't understood Rich's last proposal (item
	2 of his mail from 13 Sep 95). I don't know how software shall
	process it to check the tree automatically for conformance, or
	do installation. And for me, that's the bottom line -- a
	registry that does not provide informations to handle these
	tasks is not my beef.

    [Rich, Joachim, sebastian; -- Tobi Oetger [sp?]?!]

 -- Reserve package name local/, FWIW.

	I'm not convinced that it will really help, but it won't
	introduce problems either; so we should just add it.

    [Alan, Karl, sebastian]

Joachim Schrod			Email:
Computer Science Department
Technical University of Darmstadt, Germany