[REBOL] Re: Who read this mailing list?
From: santilli:gabriele:g:mail at: 25-Nov-2007 17:14
2007/11/24, Giuseppe Chillemi <gchillemi-aliceposta.it>:
> The new VID is on only on Carl's mind
That's not really correct. I think it's 90% done (though we're going
to re-evaluate the design soon, and I may need to rewrite some parts),
even though the remaining 10% is the one that takes the longest (the
details)... so it would probably need 3-4 weeks of work (which can
mean two months real time because I may also be doing other things in
the meantime, like protocols, testing, R/S on R3, and so on). (Keep in
mind that 90% required less than 3 weeks, done over about a month - we
started working on VID in mid July and I had to go back to Qtask in
mid August.)
> In this situation Rebol3/Core will need at least 3 months to leave the
> pre-alpha stage and go alpha.
That may be true as there are many things that still need to be done
(tasks etc), however most of these features are not that critical, so
you can call it alpha even if tasks are not 100% done and so on. I
think it already qualifies as "alpha" since you can already write
useful apps with it (and I did).
> Then will come the time for the other
> protocols to be implemented and debugged. This mean another 5-7 months.
It can be done in parallel. Carl works on the "kernel" (the actual
interpreter, the datatypes, etc.), I work on VID and protocols (with
many others contributing mezz code, eg. thru DevBase), Richard works
on View and so on.
> Thinking in parallel, VID will be structured and engineered during this
> time. A first implementation of the new VID will start after core with
> protocols will go beta.
That's not correct - we're at the second implementation already and
don't think the third will take much (I think the current code is
flexible enough for the changes Carl wants to make, so there's not
much to rewrite. It was asked that we have "css like skinning", and
although I don't think we need all the complexity of "css like", we
can do something like that by just using the way VID3 handles style
options - you can write button "Click" options [rounding: 6] for
example, so we can just make the options names more "css like" and
there's not much more code to write).
> Should we think that 1 month is enough for a full GUI system ? No, it will
No, about 2 are needed, but it might have been less if I was not
interrupted. Then of course we still have to refine the edges after
that, and maybe it will take 6 months or so to call it "release
candidate", but it's not that hard. It's REBOL we're talking about,
not Java. :)
> 2010-2011 Seems the time-frame for the firsts maturity and usabilty of
> Rebol3 in big projects.
It may be true that it will take a few years for it to be as mature as
R2 is; but, keep in mind R2 was done in like 6 months, and about one
year later we got View (2000). In 2001 I was already writing
commercial applications with it - some are still in use today even
though they run in View 1.2.10 or so. Even though RT may be much
smaller now, I think you are a bit too pessimistic. OTOH, since many
things will be left to the community, it's hard to predict when we'll
have a "critical mass" of modules and plugins available.
Anyway, that was a good post, and I completely agree with the idea -
there's no need to wait for R3, and even if it's going to take some 3
years more, the world won't fall in the meantime.
Regards,
Gabriele.