[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Math font announcement in TTN 4,2



(This is a late reply to Frank's statement of Sat, 21 Oct 1995. 
Just like him I was busy with `real work' last week and didn't 
have time for a well-thought-out reply earlier.)

Frank,

first I would like to thank you for your statement clarifying the
position of the LaTeX3 project on future math fonts work.  This was 
more or less the kind of statement I was hoping to get. At least we 
are all up-to-date now about the state of affairs, even those who
(like me) haven't been reading LATEX-L so far. (But maybe I should?)

Now, let me comment on some specific points:

I wrote:
> > Nevertheless, I want to express my concern about the way things have
> > been handled here.  In particular, let me emphassize that I consider
> > it very important that any ongoing or future work in this area will 
> > be open and transparent to members of this working group, especially
> > since there are still a number of important design decisions awaiting
> > to be tackled when preparing a prototype implementation.

Frank:
> I beg to be allowed to disagree. 
>
> I think that the work done by Justin (with the help of a lot of people
> on the math-font-discussion list) gives a good basis for actually
> implementing a prototype. I don't think it would be helpful to redo
> the analysis before having that prototype available. In other words, I
> think that it was time and still is to implement exactly what we have
> got so far. If the ltx3 project would be spending money on that it
> would be that intermediate goal and not some long term future TWG
> goal.

OK, I can understand your position from the point of view of providing
resources and spending money on the project.  Also, I didn't mean to 
suggest redoing the analysis or reopening the discussion on everything.  
Justin's report is certainly a good starting point, but I'm unconvinced 
that it is detailed enough for building a prototype implementation.  
What we've got so far (apart from various helpful technical studies)
is basically just a sketchy list of symbols and a proposal how they
should be distributed among the various encoding tables.  However, 
we still don't have anything like specific font tables and, even 
worse, there are some cases where the number of symbols considered 
for inclusion into a font table exceeds the magic number of 256.

Sooner or later some design decisions will inevitably have to be
taken, and these are the important decisions where I would like to 
see this group being consulted so that people from various fields 
(math, physics, chemistry, computer science, whatever) have a chance 
to comment on it in order to achieve the best-possible compromise.  
These are fundamental decisions which I wouldn't like to see being 
taken alone and possibly in haste by the individual working on the 
actual implementation since the result -- even if intended to be
only as a prototype -- may end up as a quasi-standard for everyone.

Remember the story of the very first implementation of TeX in 1977
outlined in Knuth's paper ``The Errors of TeX''.  Knuth wrote up 
what he thought was a pretty complete specification and handed it 
to some of his students to implement while he was traveling abroad.  
When he returned they implemented something what he found wasn't
general enough and he decided to do it himself.  It was only then 
that he realized how much was left unspecified in his original 
specification, facing important design decisions every few minutes.

I'm somewhat afraid that we might end up in a similar situation 
if someone were hired to produce a prototype implementation based 
only on Justin's report without further consulting this group, 
but maybe I'm a pessimist in this case.

> I think the fact that the work of the TWG seems to have come to an end
> after the ltx3 project stopped putting resources into it confirms the
> analysis that for getting results that would be the next step.

I basically agree since unfortunately no-one here seems to have 
enough time to get something going.  Just in case this wasn't clear: 
I don't object the plan to produce a protoype implementation by the 
LaTeX3 project, but one may have different opinions about the `how'.
(Besides, in a private reply to me someone expressed his worries 
about what he called `the LaTeX-only doctrine'.  I don't want to 
comment on that view, but it shows that it's not only me who is
worried one way or another about the `how'.)

>   even if this report was unfortunately not announced to this list it
> was available to the head of this working group with requests for
> comments. I know that we all are busy but given the long time in which
> nothing happened since the report got public also suggests that it is
> good enough to be put on as a prototype implementation.

As mentioned above, I'm unconvinced that it is complete enough which
doesn't mean that it is severly deficient but simply that more work 
is needed. It is exaclty the lack of time that's the problem here. 
For instance, I always wanted to cross-check all this symbol counting 
in Justin's report, but never got beyond reading and filing it.

> and excuse me, but i guess i have another comment concerning TWG's
> openness and design decisions: 
[...]

Without going into details let's simply say that I don't share your
impression of TWG's in general being unproductive.  I can't speak 
for many TWG's, but at least my experience from TWG-TDS is that TWG's 
can very well be open and productive, even though it may still have 
taken a year from the start to first public-review draft and another 
half year to the final version.  I wish the same would apply here.

> A similar remark applies to the math font TWG. I would be happy if the
> ltx3 project wouldn't need to put in any resources into that area as i
> don't think it lies at the heart of what we are trying to achieve. But
> since basically nothing happened since Cork to produce a standard for
> math fonts that would complement the Cork standard, we finally decided
> to put some resources in from our side.  Which in this case meant
> having a student working on this project.  During that time the TWG
> was active as well (perhaps triggered into action?) but that stopped
> more or less with the end of Justin's assignment. 

You forget that the TWG (and this mailing list) was formed only after
the Aston TUG meeting in 1993 during the time of Justin's assignment.
Of course, it is unfortunate that discussions stopped rather abruptly 
after Justin stopped coordinating it, but you can't blame this TWG that 
nothing happened before, simply because it didn't exist at that time. 

> So after having
> another <void> for more than a year from the general TeX community I
> made up my mind that we (that is the latex3 project) do need at least
> the prototype.  But since our resources are very limited I'm not
> prepared to put them into another design phase but I would put them
> into implementing that prototype and then see if it is good enough or
> if it needs improvements. In other words i guess we would be prepared
> to put some money from the royalties of the LaTeX Companion into that
> effort but it would need to be a project will clear deliverables which
> are a usable prototype.

OK, again I can understand your point of view since any prototype
produced either in open or closed development is better than nothing.
I'm looking forward towards this prototype, but don't be surprised 
if you stumble across unresolved design decisions again and again 
during the implementation process.

Consider this a well-meaning warning, not a criticism or flame.
Thanks again for your clarifying statement.

Cheers, 
Ulrik.