[REBOL] Re: Contribute to REBOL/Core
From: petr:krenzelok:trz:cz at: 5-Nov-2000 11:06
----- Original Message -----
From: Dan Stevens <[dan--rebol--com]>
To: <[ally-list--rebol--com]>; <[rebol-list--rebol--com]>; <[command-list--rebol--com]>
Sent: Thursday, November 02, 2000 12:12 AM
Subject: [REBOL] Contribute to REBOL/Core
> Dear REBOL Community,
>
> For those of you "chomping at the bit" to contribute fixes, new functions
> and other improvements to REBOL/Core, we've added a process for reviewing
> and accepting your contributions. We sincerely appreciate these
> contributions and attempts to further the REBOLution. We will do our best
> to review them on a timely basis and determine if they can be added to the
> official REBOL/Core releases.
>
> Please go to http://www.rebol.com/contribution_source.html for more
details.
> Ingo, Pekr, Andrew, Allen, Elan, and the many I've left out...please,
have
> at it
! We look forward to your additions...Dan
OK, I just returned from my trip to Prague, where I visited very nice
concert of Finish band called NightWish and reading thru tones of emails I
finally managed to answer this one ....
So Dan, I will describe situation as I see it and I would be glad if you
could regard my comments as just my POV, nothing more .... :-)
As I am person always trying to see things from bigger perspective, I am not
sure I can help REBOL in the level required. Reading thru suggested link I
noticed you are accepting contributions at mezaninne level.
1)
... as for mezannine level, I will probably let the role of contributors
free for other, much more advanced rebollers as Andrew, Gabriele, Ladislav
and many others surely are :-)
Well, there is just very few mezannines I use, as I know they could bring
problems to other rebollers not having them around. I can see current RT's
aproach similar to earlier suggestion of creating some standard, by most
users agreed upon - free source library. The only one disctinction seems to
be inclusion of suggested mezzanines into REBOL environment ...
I currently use:
... copy some more parts of VID styles. e.g. changing the progress bar
setting is reflected thru all progress bars in display ...
time-of: func [date [date!]][fourth date]
words-of: func [object [object!]][next first object]
values-of: func [object [object!]][next second object]
... as users are not required to know details of rebol object implementation
that way ....
2)
the only one valueable advancement I can see for REBOL is advancement of
Core technology, which is nowadays beeing postponed in favor of product
development (/Express). But I am not surely alone thinking the priorities
are exchanged that way. If RT has to bring some product to life because VCs
want it that way, we can tolerate it, but it doesn't mean we will agree
here.
My master Yoda always tells me - reality is good reflection of your
advancement ;-) And the reality for REBOL as I can see it is as follows:
- although planned for some future, not componentised properly, which is
reflected in following:
- many versions of Core around the product lines ...
- Apache based on Core, no way how to plug library component ....
Command without View capabilities ... Command containing possibly unwanted
components (Oracle etc.) ... I can see per-component pricing as the BIG win
for all the parties
- Elan and others expressed their practical needs - View with Command,
Apache with mySQL, etc. we have first signals REBOL developers have to
switch to other solutions ...
- a little bit underpowered capabilities of /View. We are near to enter some
interesting areas of development but stopped half-way. Single primitives for
drawing, filling areas could help. Bo's dialect for drawing is fine, but if
even my 650MHz Duron can't help the situation, something is wrong here (I
don't remember who's the author of demo of moving boxes connected by lines
...)
3) suggestions ...
a) long term
- improvements to parser as suggested on Robert's site ...
- introducing modules and dynamic components ASAP ...
- ability to integrate external windows in View one. e.g. VID panel
displaying mpeg player output
- more orientation towards media, presentation, entertainment market ...
b) short term
- from beta 1 View I suggested to implement change of state of mouse
pointer, forbidden mouse pointer (replaced by two nested faces) - nice for
some kiosk styles. I also suggested transparency at top level, but not sure
it's technically easily doable
- /files, /dirs, /dir-tree or something like that refinement to 'load to
speed up the work with directories. If we want to get directories and files
separately, we have to iterate thru block of read items ... I know it's
special case, but well, even Holger said files are special case :-) Would
turn file requestors easily usable even on larger directories ...
- /deep refinement to find. I am not sure if others would find it
interesting, but it's kind of functionality turning nested block structure
into one level, flat, block, at least during search ...
Suggestion to myself - patience ;-) Well, actually - I just hope /Express is
going to be great product, because if it is not - we just lost some year of
possible advancements being made to REBOL technology in general ...
So folks,
Cheers,
-pekr-
