New TIP 279 on Tcl OO core facility has similar goals to #257

Roy Terry

unread,

Oct 9, 2006, 3:22:12 PM10/9/06

to

Just so people can be informed. There is
some lively discussion about the details of
implementation on TIP 257 OO core for tcl.

Gustaff Nuemann, a developer of XOTcl,
has authored an alternative TIP 279 and has a demo
implementation in progress (perhaps
available now). Some discussion has been
made on the Tcl core mailing list.

http://www.tcl.tk/cgi-bin/tct/tip/279.html

Robert Hicks

unread,

Oct 9, 2006, 3:41:49 PM10/9/06

to

I will be curious to see whether TIP 279 is "accepted" as a TIP.

Robert

GN

unread,

Oct 10, 2006, 3:08:22 AM10/10/06

to

Roy Terry schrieb:

> Just so people can be informed. There is
> some lively discussion about the details of
> implementation on TIP 257 OO core for tcl.
>
> Gustaff Nuemann, a developer of XOTcl,
> has authored an alternative TIP 279 and has a demo
> implementation in progress (perhaps
> available now). Some discussion has been
> made on the Tcl core mailing list.

As mentioned in the tip, here is an implementation based on the current
xotcl snapshot (from Oct 5) available. For a definition of TclCOOL (Tcl
Core Object Oriented Language) based on this, see:
http://groups.google.at/group/comp.lang.tcl/browse_frm/thread/993ece6cae9cbe22/bac4d429b2269cd2?lnk=st&q=xotcl&rnum=1&hl=de#bac4d429b2269cd2

Larry W. Virden

unread,

Oct 10, 2006, 9:58:38 AM10/10/06

to

Robert Hicks wrote:
>
> I will be curious to see whether TIP 279 is "accepted" as a TIP.

Well, if you mean, is it truly a tip - yes. If you mean "has it been
voted on and approved", no, neither of the tips have yet gotten a vote.

If there are people here who want Tcl to be support object oriented
programming in the core, and you have a reason for wanting this other
than "people make fun of tcl because it doesn't have oo in the core"...
i.e. you have technical reasons why you need OO, you are strongly
encouraged to read up on these two tips, as well as the tip that was
already accepted several years ago, but awaits someone to actually do
the work, relating to OO and to provide input.

Otherwise, you may end up with something that one person thinks is OO
and someone else doesn't...

Robert Hicks

unread,

Oct 10, 2006, 10:13:50 AM10/10/06

to

Actually, I fall into the "nice" but I can use Snit or XOTcl as is. So
anything that helps Snit (and to a lesser extent XOTcl) is fine with
me. : )

I was actually thinking that a TIP once created had to be "accepted" by
the TCT. I guess that is not so.

Robert

Don Porter

unread,

Oct 10, 2006, 10:38:37 AM10/10/06

to

Larry W. Virden wrote:

> Otherwise, you may end up with something that one person thinks is OO
> and someone else doesn't...

That's going to happen no matter what we do.

--
| Don Porter Mathematical and Computational Sciences Division |
| donald...@nist.gov Information Technology Laboratory |
| http://math.nist.gov/~DPorter/ NIST |
|______________________________________________________________________|

Larry W. Virden

unread,

Oct 10, 2006, 11:45:47 AM10/10/06

to

Don Porter wrote:
> Larry W. Virden wrote:
>
> > Otherwise, you may end up with something that one person thinks is OO
> > and someone else doesn't...
>
> That's going to happen no matter what we do.

Good point... well, it wouldn't happen if we didn't put anything OO
into Tcl... then we'd just maintain the status quo of people whining
about not having OO, even when they don't even agree on what that would
look like...

Larry W. Virden

unread,

Oct 10, 2006, 11:47:32 AM10/10/06

to

Robert Hicks wrote:
> I was actually thinking that a TIP once created had to be "accepted" by
> the TCT. I guess that is not so.

I believe that "accepted" status that you see on TIPs means that
discussion and a vote has occured. - that someone now needs to move
forward, write code, work in the cvs to add the code, write doc and
test cases, etc.

Ulrich Schöbel

unread,

Oct 10, 2006, 7:55:01 PM10/10/06

to

QOTW candidate!

Donal K. Fellows

unread,

Oct 17, 2006, 5:23:16 AM10/17/06

to

Robert Hicks wrote:
> I will be curious to see whether TIP 279 is "accepted" as a TIP.

The submission met the criteria for being a TIP (i.e. it is a proposal
to improve the Tcl core, and it has a suitable copyright declaration)
and did not constitute a duplicate of an existing TIP. Given that, I
assigned a TIP number and published it; that's what I do as TIP Editor.

It's no secret that I prefer #257 though. :-D

Donal.

GN

unread,

Oct 23, 2006, 2:55:24 PM10/23/06

to

In case someone is following the development of TIP #279
(http://www.tcl.tk/cgi-bin/tct/tip/279.html), here is an implementation
of an subset of itcl that runs already parts of the itcl regression
test (eg all 60 tests of protection.test, the 58 tests of basic.test,
three with different error messages). On methcall benchmark from the
oo-benchmarks, it is about 20-30 % faster than c-coded itcl 3.3.

itcl implementation
http://media.wu-wien.ac.at/download/itcl.tcl

based on the current xotcl snapshot

http://media.wu-wien.ac.at/download/xotcl-1.5.3-alpha2.tar.gz

my itcl test suite
http://media.wu-wien.ac.at/download/mini-itcl.tcl