RE: Another BOF idea

Carl,

Please don't take this the wrong way, but I wonder how many people are
actually trying to use Pollock for their applications given that its nowhere
near completion AFAIK? Isn't it a bit premature to spend time addressing the
problem knowing that framework may change shape many times over in the next
year or two? Perhaps I'm underestimating the current impact of this problem
on the community as I thought that people were just experimenting with
Pollock to see what its all about, in which case performance issues would be
of least worry for them and they'd sleep well knowing that performance
optimization is still on Sames' radar and will not be overlooked.

One other thing to consider is the fact that officially all BOF sessions
will be scheduled from 5:30PM to 6:30PM and most of them on Tuesday, so a
lot of tough choices are going to be made. So far Packaging and GemStone
BOFs have been confirmed.

Cheers!

-Boris

--
+1.604.689.0322
DeepCove Labs Ltd.
4th floor 595 Howe Street
Vancouver, Canada V6C 2T5

[hidden email]

CONFIDENTIALITY NOTICE

This email is intended only for the persons named in the message
header. Unless otherwise indicated, it contains information that is
private and confidential. If you have received it in error, please
notify the sender and delete the entire message including any
attachments.

Thank you.

-----Original Message-----
From: Carl Gundel [mailto:[hidden email]]
Sent: Wednesday, April 05, 2006 11:12 AM
To: [hidden email]
Subject: Re: Another BOF idea Re: [Ann] BodyBuilder packaging suite

Heh.  Well, I was thinking that probably a lot could be discovered by
profiling the known case of inserting text into a TextEdit control that has
been added to a form.  I spent some time looking over this myself but
haven't nailed it yet.  I notice that it spends a LOT of time invalidating
rectangles and redisplaying things.  It wasn't clear to me whether these
operations themselves are slow (seems unlikely) or that perhaps the whole
process is iterating out of control.
 
I hope that several pairs of eyes might do better.
 
So, I wonder who else here is interested in a Pollock Performance BOF?

        ----- Original Message -----
        From: Charles A. Monteiro <mailto:[hidden email]>  
        To: Carl Gundel <mailto:[hidden email]>  ; [hidden email]
        Sent: Wednesday, April 05, 2006 1:21 PM
        Subject: Re: Another BOF idea Re: [Ann] BodyBuilder packaging suite

        Just don't run the sunits at the beginning of the BOF :)
       
        On Wed, 05 Apr 2006 11:21:51 -0400, Carl Gundel
<[hidden email]>  
        wrote:
       
        > Another idea a had for a BOF I had is for Pollock performance.  To
get a
        > few interested people together to do as Sam suggested and run the
        > performance tools on Pollock to nail down some bottlenecks and
maybe fix
        > one or two of them.

...... Original Message .......
On Thu, 6 Apr 2006 11:30:59 -0700 "Boris Popov" <[hidden email]>
wrote:
>Carl,
>
>Please don't take this the wrong way, but I wonder how many people are
>actually trying to use Pollock for their applications

Well, I am using Pollock for my application.

> given that its nowhere
>near completion AFAIK? Isn't it a bit premature to spend time addressing
the
>problem knowing that framework may change shape many times over in the next
>year or two?

I need to solve this problem. If I discover some way to speed Pollock up,
I'll forward it to Sam.  Maybe he'll find it helpful.

>Perhaps I'm underestimating the current impact of this problem
>on the community as I thought that people were just experimenting with
>Pollock to see what its all about, in which case performance issues would
be
>of least worry for them and they'd sleep well knowing that performance
>optimization is still on Sames' radar and will not be overlooked.

Sam said at the last Smalltalk solutions that Pollock was "finished"  (or
done, or ready, or some such word) and encouraged us to use it.  So I did
just that.  On the one hand Pollock works pretty well, but there are
problems.  Performance is one of them.  

>One other thing to consider is the fact that officially all BOF sessions
>will be scheduled from 5:30PM to 6:30PM and most of them on Tuesday, so a
>lot of tough choices are going to be made. So far Packaging and GemStone
>BOFs have been confirmed.

That's fine.  It was just a suggestion.  Perhaps it would more appropriate
to just try and grab a few interested people and sit down together to work
on this.  :-)

Regards,

-Carl

>Cheers!
>
>-Boris

> ...... Original Message .......
> On Thu, 6 Apr 2006 11:30:59 -0700 "Boris Popov" <[hidden email]>
> wrote:
>>Carl,

>>Please don't take this the wrong way, but I wonder how many people are
>>actually trying to use Pollock for their applications

> Well, I am using Pollock for my application.

WithStyle V4 is completely Pollock based. (psst, it ain't slow for us!)

>> given that its nowhere
>>near completion AFAIK? Isn't it a bit premature to spend time addressing
> the
>>problem knowing that framework may change shape many times over in the next
>>year or two?

You know, it really won't change that much. Sam just says that are a disclaimer so he can go and change things.

Cheers,
Michael

Oh c'mon. You can't really know whether it'll change or not and if yes how
much. That's just a guess. Not even Sames does and that's the convenience of
having an unreleased product. While I understand your reasoning about the
framework having been through enough brainstorms to be considered solid,
even slightest refactoring could change its performance profile quite a bit
and therefore invalidate efforts of an outside group of people trying to
address current issues. If anything, all changes coming out of the effort
would have to be integrated into the main tree fairly quickly to not fall
prey to entropy, but you'd have to ask Sames whether he is willing and has
time to do that right now before he even gets into performance optimization
stage himself.

Cheers!

-Boris

--
+1.604.689.0322
DeepCove Labs Ltd.
4th floor 595 Howe Street
Vancouver, Canada V6C 2T5

[hidden email]

CONFIDENTIALITY NOTICE

This email is intended only for the persons named in the message
header. Unless otherwise indicated, it contains information that is
private and confidential. If you have received it in error, please
notify the sender and delete the entire message including any
attachments.

Thank you.

-----Original Message-----
From: Michael Lucas-Smith [mailto:[hidden email]]

Sent: Thursday, April 06, 2006 3:21 PM
To: Carl Gundel
Cc: Boris Popov; [hidden email]
Subject: Re[2]: Another BOF idea



> ...... Original Message .......
> On Thu, 6 Apr 2006 11:30:59 -0700 "Boris Popov" <[hidden email]>
> wrote:
>>Carl,

>>Please don't take this the wrong way, but I wonder how many people are
>>actually trying to use Pollock for their applications

> Well, I am using Pollock for my application.

WithStyle V4 is completely Pollock based. (psst, it ain't slow for us!)

>> given that its nowhere
>>near completion AFAIK? Isn't it a bit premature to spend time addressing
> the
>>problem knowing that framework may change shape many times over in the
next
>>year or two?

You know, it really won't change that much. Sam just says that are a
disclaimer so he can go and change things.

Cheers,
Michael