Donal,

>  From my perspective,
> > it looks like you're just trying to block me, to "teach the controversy"
> > and other stuff like that. That annoys me. Don't do that.

Well, asking for opinions on a public community list bears the risk of 
getting critical echos, indeed. But be assured, I am neither in the 
position nor actively seeking to block your TIP or CFV. How could I, you 
put a lot of effort in TclOO, no question.

> > It's not my TIP, so moving it forward to being implementation-ready
> > would seem to me to be someone else's job[*]. There's a lot of working
> > code that implements #257, and it's (probably) fast and robust. 

I disagree. For two reasons:

	First, TIP#279 provides a prototypish implementation of your TIP#257 in 
terms of TclCOOL. It shows how to achieve TIP#257 sketches in a few LOC. 
As both TIPs settle at a level of being an OO-language infrastructure 
(correct me if I am wrong here), your TIP is missing this attempt 
entirely. So TIP#279 authors did something that you would have been 
expected to provide (turning the coin), if I read your statement above 
correctly?

	Second, your TIP (as I read it) is not about yet-another OO smell for 
Tcl which happens to be in the core, but a common OO language 
infrastructure, in support of language designers. If it is not, then it 
is not comparable to TIP#279 which would then be more generic. In doing 
so, it is rather important to organise for a heads-up and coordination 
at a _human_ level. And that's what you actually did, unfortunately only 
with a selected group of people. I am remembering the announcement of 
TclOO 0.2 at comp.lang.tcl which contains a line:

>  Many small features, mostly hidden out of sight, for support of
>      Arnulf Wiedemann's itcl-ng. 

Also your itcl-sprinkled in-code comments and the active discussion at 
tcl-core show that there has been this kind of coordination attempt. But 
has there ever been a personal note or invitation to XOTcl authors, i.e. 
Gustaf Neumann and Uwe Zdun, to actively participate? My personal 
impression is rather that you exhibit a kind of XOTcl-phobia for 
whatever reason.

>  admit it's a few months since I last ran performance tests, but TclOO
>  0.1 beat XOTcl 1.5.5 handily by about 2x on some simple tests (see
>  http://wiki.tcl.tk/18152 for the evidence, extracted from my paper last
>  Tcl conference). If someone were to redo and rerun[**], that would be a
>  good deed in itself.

Voilà (best out of 10; MacBook Pro 15" 2,33GHz, Mac Os X 10.4, Tcl 
8.5.2, TclOO 0.2, XOTcl 1.6.0):

TclOO 0.2
--------------------------------------------
689257.6371297024 calls per second of: f bar
64447.488714664694 calls per second of: [foo create f] destroy

XOTcl 1.6.0
--------------------------------------------
368686.4564283765 calls per second of: f bar
88341.87365145577 calls per second of: [Foo create f] destroy

First, the superiority you reported in your tcl2007 presentation seems 
to be a moving or volatile target. Second, this kind of benchmark 
shouldn't be an argument pro or contra of #257 or #279 as they are 
implementation-specific and have nothing to do with the design of the 
language(s) which should be your focus. The only thing they tell is that 
you realised a better performing method dispatch which is certainly not 
what OO language design is about (and which can certainly be done for 
XOTcl as well)

>  In short, if there's a problem with TclOO, do speak up now!

1. I am missing documentation on your design decisions (note, not use 
docs!). TIP#257 itself reads as "how do I differ nominally" ("instproc" 
is to be called "method") but you don't provide references (scientific 
ones?). References, especially, beyond the Tcl point of view that 
inspired your thoughts and guided your decisions. i am also missing 
those in your tcl2007 contribution*.

2. Getting rid of the XOTcl-phobia and trying to get people in a 
coordinated activity. Both sides can only win, full-fledged tcl core 
developer on the one, experienced OO language designers with Tcl 
affinity on the other hand. This starts by getting in touch directly and 
settling the prevailing scepticism ...

//stefan

* 
http://www.tcl.tk/community/tcl2007/proceedings/programmingtechniques/OOforTclFellows.pdf


-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Tcl-Core mailing list
Tcl-Core@[...].net
https://lists.sourceforge.net/lists/listinfo/tcl-core