Tcl's Future was: Apache 2.0 TCL solution

Tcl's Future was: Apache 2.0 TCL solution

Post by Robert Hic » Thu, 03 Aug 2006 23:29:53



<snip>

Do you have to know "C" to help?

:Robert
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by davidnwelt » Fri, 04 Aug 2006 00:36:24

> I doubt anyone will answer. There isn't a clear leader for Tcl, or even

To me, this is the bigger problem, rather than ...


... engineering. I've learned a lot about how to write good, well
tested code from the core team.

Tcl is actually pretty robust compared to a *lot* of open source stuff
out there. I see the problem as more of a marketing/human one. For
instance, OSCON was a fun time, with all kinds of "important" free
software people. Where was Tcl? It's completely off the radar.


Sounds like 'release early, release often' needs to be applied a bit
more liberally, realizing that it probably means slightly less robust
releases.


That's a vicious cycle, because the less it gets worked on, the less it
gets used, and the fewer new people get interested, meaning no one to
replace core guys who go on to other things.

I don't think Tcl is quite at that point yet - the core team has added
people like Joe English and Miguel Sofer "lately", but it just doesn't
seem to be headed anywhere unless something unexpected happens. But
the likelyhood of that happening decreases as the language loses its
ability to attract young people with lots of time to hack on
cool/innovative ideas (stuff like Rails).

 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by davidnwelt » Fri, 04 Aug 2006 00:39:07


Not necessarily. Some initiatives that I started that require
absolutely *0* C knowledge:

*) The tutorial. It still needs work, too, but being able to include a
bit of documentation to get people started without pointing them at an
(expensive/difficult to find) book is a win.

*) Making sure every man page has an example of how to use the command
in question, including, perhaps, some common idioms. I think this is
largely done, but it's something that anyone with a bit of Tcl
knowledge can do.
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by simo » Fri, 04 Aug 2006 01:13:12

The reason I started the Tcl 8.5 Roadmap post last month was because of
similar concerns I have about 8.5 but there wasn't a lot of interest in
the topic.

IMHO 4 years is too long for 8.5 to be in development and release
should be regarded as a matter of urgency:

-> feature freeze now
-> 2 month bug-fix to first beta
-> 2 month bug-fix to final beta
-> release Nov/Dec 2006

This of course means that the cherished OO will not make it into 8.5.
Tile would make it in its current state as long as it didn't break the
existing tk codebase (which as far as I know it doesn't 'cos it's all
in the ttk namespace). 8.6 in Nov/Dec 2007 can have these features if
they're ready.

With some focus around a release timetable I'm sure the wider community
would help out as much as possible but it will need members of the core
team to be driving it along.

SImon Geard
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by Michael Sc » Fri, 04 Aug 2006 02:09:16


XXXX@XXXXX.COM schrieb:



The tutorial is nice, but as you correctly say needs some more work, to
bring it up to date and add some of the new 8.5 stuff...

This is done for Tcl, but the Tk man pages could use examples...

Michael
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by Robert Hic » Fri, 04 Aug 2006 03:50:03


"Someone" should make a page with a list of to-do items for 8.5 on the
wiki and then people can volunteer for those items.

Robert
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by Neil Madde » Fri, 04 Aug 2006 04:00:15


[...]

The sourceforge.net tracker for tcl and tktoolkit projects is the place
to look.

-- Neil
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by Robert Hic » Fri, 04 Aug 2006 06:14:42


Are those for 8.5 specifically? I think the vector is "what needs doing
to finish 8.5 sans oo"...

Robert
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by George Pet » Fri, 04 Aug 2006 07:22:12


How will we know who is working on what? Should people working on bugs
or documention just append to the report "I'm working on this?"

There's something very motivating about having a public and noticable
to-do list. I think it would be useful to be able to strike off items
and sign your name. I think it should be created, because right now you
must span multiple reports as I understand it to find what's needed.
If we had a list with links to the reports or patches for
code/documentation it would be much better.

-George
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by Robert Hic » Fri, 04 Aug 2006 10:04:02

Who decides what needs or doesn't need to be done for 8.5? I don't
think waiting for oo in the core is good...maybe 8.6 in 6 months? There
should also be a definitive list of what *has* to be done to get 8.5
out the door.
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by George Pet » Fri, 04 Aug 2006 13:24:26


I don't recall reading it. I'm interested.


I agree and 8.5 is full of *useful* features that should be out there.

* {expand} is good.
* Tile is ok
* [chan] (possibly a new feature or 2 I hope)
* extension unloading
* numerous TIPs that have been accepted and implemented.
* refactorings and speed improvements in some cases.
* the new clock subsystem (this is a huge change)
* the large integer/BigNum support via libtommath (included in Tcl).
* a new expr parser
* [apply] (anonymous procedures/lambda)

There are probably other good features I'm neglecting.


Yes, Tile has just required some minor stubs table exporting in Tk that
is compatible with 8.4 and 8.5.


That may be difficult, because half of the core team is inactive. Some
have basically quit (for now?).

-George
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by George Pet » Fri, 04 Aug 2006 13:28:44


My personal opinion as a core developer with cvs write access (but not a
core team member), is that we should have all priority 9 bugs resolved
prior to release, and as many priority 8 bugs as possible. But as I
noted on the wiki today we need all bugs fixed that are possible to fix.

That means documentation bugs, and anything else you can think of too.
You don't have to hack C to work on Tcl.

-George
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by George Pet » Fri, 04 Aug 2006 14:13:29


Let's get to the heart of the matter. Why hasn't the TCT focused on
this? Why isn't the TCT releasing status updates?

To most outsiders Tcl must appear just like any other old language, but
I think it's more.


I have too. I've also learned some things I'd never do, because some of
the code is or was a mess. But, I appreciate the messes sometimes just
as much, because Tcl is free, and provides quite a lot.


We don't need marketing. We need people creating more competitive
applications, and wanting to show them off.

Tk doesn't compete with Gtk+ and Qt in terms of image manipulation speed
(a common requirement today). I've had to write my own extensions for
this, and now I'm gradually porting a display widget to Windows.

Tk doesn't compete when manipulating images (think of the GIMP in Tk),
and until Tile comes with Tk we won't have themes. Even then,
applications will have to be gradually rewritten for Tile, because you
can't do namespace import -force ::ttk::* anymore. We also don't have a
standard 3D binding for OpenGL like many other toolkits. We've got a
choice of 3 or 4 different OpenGL interfaces. Good luck if you can
build some of them on non-Linux platforms (I know, because I've tried).

[snip]

I don't think it means less robustness overall. The TCT isn't doing
code reviews or being proactive about hunting down *possible* bugs from
what I've seen. Instead the TCT is reactive.


You've already given up. :( You had some hope last year, what's
changed? I've seen videos for Ruby on Rails, and I don't see the
appeal.


-George
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by simo » Fri, 04 Aug 2006 19:09:57

There have been many references to the TCT in this post. Could we
please have a statement from the someone in the CT that represents
their (collective) view of the release timetable of 8.5. The community
seem eager for it but without some drive from the centre I don't see it
ever happening.

Simon Geard
 
 
 

Tcl's Future was: Apache 2.0 TCL solution

Post by Robert Hic » Fri, 04 Aug 2006 21:32:14


I am more than willing to help where I can.

Robert