Re: [Texmacs-dev] Arch mirror (was: EPS bounding box bug)

[Joris van der Hoeven]

On Tue, 30 Sep 2003, David Allouche wrote:
> On Tue, Sep 30, 2003 at 11:51:21AM +0200, Joris van der Hoeven wrote:
> > On Tue, 30 Sep 2003, David Allouche wrote:
> > > There are unconfirmed rumors that at some point in the future Joris
> > > might try arch.
> >
> > In fact, I have still been thinking about all this and discussing
> > these issues with people here in germany. I feel that something
> > we *really need* is a more abstract tool (probably in C++ and maybe
> > even in the core of TeXmacs) for computing structured diffs and
> > computations with structured patches. This tool should be used both
> > for filesystems (which can be seen as documents) and for usual documents,
> > and especially for our future undo/redo system (in fact it will
> > really be necessary).
>
> Some well know dutch programmer and mathematician once said:
>   "In an ideal world we have everything"
> But we are not in an ideal world.
>
> Speculation about ideal tools is a vast topic (and Arch people will be
> interested by your ideas) but workflow is not improved by non-existent
> ideal tools.
>
> There are existing tools now and here which can help a lot. From my
> experience, we need something working like Arch _now_. Not something
> hypothetically working better at some point in the future.
>
> Your ideas are probably very interesting but they are not relevant
> in this discussion, which is about working tools now.

I do not agree, because the thing I want is not for the far future,
but for next year: the new data model will require it in order to
correctly handle undo/redo operations. I prefer not to learn
new tools if I know that I am going to change them within one year.
Moreover, it is very possible that there will be a student working on this,
which is increasing the probability that things get done in time.