Towards 3.9.1 . Away from What?

Towards 3.9.1 . Away from What?

I watch all of Keith's efforts with wonder and some
trepidation.

He has very bold plans and when kept after (by others)
may do an reasonable job of coding.

He is also off acting on his own on what he thinks is
needed.

I see what he is doing as undermining the efforts of
the 3.10 team.  And I wonder why this is.

Is the November revolution happening in May?

What he is proposing is creating a branch of Squeak
under his release control and to his taste.

This is something anyone has a the ability to do in a
open software enviornment.

He is appropriating the version number 3.9.1.  Which
portrays an intention for his release to be a
successor to 3.9 but close to it.

Now I question whether an independent developer of
relatively short standing in the squeak community
should have the right to give an official sounding
designation to his release.

Shouldn't the version numbers of squeak belong to the
squeak foundation and squeak board?
In other words it comes close to trademark
infringement.

Keith has the power to do what he will.  But calling
it 3.9.1 implies a blessing for his actions that he
has not formally asked for nor formally received. Some
controls should be put on what he calls it less his
efforts be taken (by those outside) as authorized and
encourgaged by the offical board.

Or he should be given that authorization by the board
to go ahead.  Right now the situation is like having a
loose canon aboard a ship.  And people should look at
it in that manner.

Yours in curiosity and service, --Jerome Peace (Bug
Tracker)










__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 

Jerome Peace <peace_the_dreamer <at> yahoo.com> writes:

> I watch all of Keith's efforts with wonder and some
> trepidation.
>
> He has very bold plans and when kept after (by others)
> may do an reasonable job of coding.
>
> He is also off acting on his own on what he thinks is
> needed.

>From the point of view of the dev list, it doesn't seem to be completely
striking out on his own: the "towards 3.9.1" thread seemed broadly
supportive. Edgar quite rightly put forward questions to do with future
developments in the same thread, but I didn't see those as negative, and
support his vision and reasons for raising the questions.
 
> I see what he is doing as undermining the efforts of
> the 3.10 team.  And I wonder why this is.

Do you mean you wonder why you see it this way, or do you mean you wonder
why it is undermining the 3.10 team?
 
This is a fair point. IMO if used "officially" the 3.9.1 name should indeed
be used for a release that is "offically blessed" by the community.

I do not know whether Keith thinks of it in his mind as an official
release. Perhaps he does (I'm sure somebody would tell me if so). Perhaps
he refers to it as 3.9.1 simply because that is the closest traditional
version name that represents the state he's trying to get an image to.

Either way, the reason I think his effort could be worthwhile in addition
to the 3.10 team (I in no way wish to undermine the 3.10 team's efforts) is
because it will get the fixes and changes he's incorporating "into the
wild" that much sooner.

Admittedly, if it's not an official 3.9.1 release it's likely that not as
many people will use Keith's resultant image than if it was official.
Despite this, if his efforts are publically available, some people are
going to use it.

It has struck me for some time that Mantis and posts to the list together
contain a *lot* of useful fixes and suggestions that spend a long time
there, not being incorporated into a downloadable image. Many of the fixes
are very important. I submit that while it is easy for experienced
Squeakers to file/MC these fixes into their images, or incorporate them
into their own dev environments, it is not easy for somebody new to do so.

I think the 3.10 effort would benefit from seeing some of these fixes it
may wish to use incorporated into an interim effort with public visibility:
this will give real user feedback on whether the fixes are viable or not,
without the need for a lengthy 3.10alpha/beta (at least for those fixes),
and importantly exposure to users whose relative lack of Squeak experience
may well lead to feedback on those fixes that those more skilled with
Squeak may not have found.

Certainly in my experience it is those new to a system who expose its
faults or shortcomings the quickest, and often the same people who have
great ideas for something new directly because of their lack of experience.

Of course the same is true for fixes Keith may choose that do not tally
with those the 3.10 team would choose. Here I think that all parties would
benefit from plenty of communication to try and make sure the two efforts
do not clash.

Early and regular release of software (providing it is ready for release!)
is good, but not at the expense of a divergent community.

Just my (overly long) thoughts.

Cheers,
Simon

El 5/6/07 6:24 PM, "Simon Kirk" <[hidden email]> escribió:


Ralph put some pages into swiki http://wiki.squeak.org/squeak/5919 " The
3.10 release " for any could see what we try to do.
See also http://wiki.squeak.org/squeak/5934 "List of Mantis issues ready to
be included in 3.10"
I dig into Mantis and put some what I think fills the requirements.
If I don't put more fixes/enhancements is for :

1) No test
2) Some question to report by a master Squeaker
3) The reporter is not answering some question what I send from Mantis.
4) Problems with Monticello what I could't solve until now. The actual
process requires what all should be of the form:

A nnnn.cs with only a Postscript what tell what .mcd  or .mcz the image
should load of 3.10 repository goes to ftp .
The user only need push the "load code updates" button of the Squeak flap.

The image after this load should have all test green.

Edgar

Hi Edgar

This process sounds great. I never meant to say that what you're doing is
not visible, apologies if I gave that impression. By visibility I meant the
code would be available in an image so that a user would get first-hand
exposure to it.

By "available in an image" I mean an image that a user can download with
the changes already loaded. From the impression I had gathered before, this
didn't seem to be easily available.

Now I've been pointed to the relevant pages in the wiki I see a lot of very  
useful information, and I see that things are well documented and easy to
get. Looking at www.squeak.org, I'd say that some of this information from
the wiki should be fed through onto the website: It can't hurt to show that
the Squeak Community is actively developing their environment in as many
ways as possible.

I'm a little worried though that I was so ignorant to what was going on
with 3.10, but I'm ready to admit that was down to my not having read the
list/website/weekly squeak closely enough to know. Still, I do read the
list almost every day, so I hope my lack of knowledge isn't an example of
others'. Let's hope it was just my blindness :)

Simon

>
> Towards 3.9.1 . Away from What?
>
> I watch all of Keith's efforts with wonder and some
> trepidation.
>
>  
wonder I like!
> He has very bold plans and when kept after (by others)
> may do an reasonable job of coding.
>
> He is also off acting on his own on what he thinks is
> needed.
>  
Not really true, someone said 3.10 is about process, not features. The
'process' element of 3.10 has had no discussion, specification or
roadmapping! 3.10 Itself has been progressing under Edgars 'expert, but
manual' direction, which I have been monitoring and supporting.

I actually feel undermined a bit by 3.10, since 3.10 was announced as
being about process, but has continued without actually waiting
for/investing in that process development. (at least, I should say
publically, there may be stuff going on behind the scenes of which I am
not aware.)

I myself have had a vision for the process side of things there is
nothing revolutionary about it, just pragmatic.

The initial "proposed" 3.9.1 bug fix scripts have been sitting on that
"experimental" website for several months while I developed
Rio/Filename/Whatever. If they dont get used now, then they may not get
used, seems silly to waste them. Many of these fixes are in 3.10 already.

I have been working on some process elements which make perfect logical
sense and are needed, and their need has been expressed by Ralph on
occasion: For example I have wriitten a non GUI TestRunner, which may be
triggered from the commandline, and monitored remotely via a web
browser. It could be useful for non GUI KernelImages too.

I extended SUnit to allow tests to be flagged as for a particular
platform or release, and appropriately filtered and to enable many
different ways of specifying test suites for different purposes. None of
these are my own ideas, these are things people have requested over the
years. Some of these features are intended to enable integration with
SSpec an alternative to SUnit.

If you are going to regression test an image on multiple platforms and
vms etc then you simply need this feature. If you are going to
regression test a package on multiple platforms, vms and images then you
will also desire this feature and a test runner that works in as many
images as possible, manageable by some externel test running
program/script. TestReporter may be that runner. Adding these features
to SUnit its hardly the act of a rebel.
> I see what he is doing as undermining the efforts of
> the 3.10 team.  And I wonder why this is.
>
>  
Undermining 3.10? I added these improvements to TestRunner back in
september last year!

The 3.10 team has not been working on process much at all, apart from
the Monticello atomic loading fix.

The 3.10 guys have already taken the 3.9.1 bug fix list as was as a
starting point.
> Is the November revolution happening in May?
>
>  
What november revolution, is there a plan?

When did 3.8.1 come out? Just after 3.9 if I recall. I might be wrong
about that though.
> What he is proposing is creating a branch of Squeak
> under his release control and to his taste.
>
>  
Actually I have been talking with Stephane all along about all sorts of
aspects of this process. Stephane has held back, in order to let 3.10
roll. But if he would like to take the opportunity to help put some
icing on his 3.9 cake then why not. Question away: the way I see it is I have no more or less right than
anyone else, until someone objects and it goes political, or someone
else has a better idea.
> Shouldn't the version numbers of squeak belong to the
> squeak foundation and squeak board?
> In other words it comes close to trademark
> infringement.
>  
Read the website it says, "This is the site for collaborating on the
specification for 3.9.1 (Pending objections, this is the de-facto 3.9.1,
the future of squeak 3.9)"

First, please note the word COLLABORATING, its not about what I want in
3.9.1. It is and always has been something I have attempted to
facilitate the community to engage in. The fact that few others have
actually engaged in it, (and I have invited them) is not my fault, nor
does it make me a loner at heart, just a loner by default.

The "Essential Fixes" page are some of what I need to support this
process that is used to generate 3.9.1 itself from this website.

You may think I am being rather forward in saying "this is the de-facto
3.9.1", but this bug fixed version has been sitting on that website for
2/3 months now and nothing has been happening to make it happen. It
needs eaither a) Someone to push it through, to do it and make it
official, or b) Somone to object and can the idea altogether so that it
becomes a mini-fork. I left both options open in the above statement. I
am being more forward now to provike action one way or the other.

So far we have a) Stephane and b) Jerome
> Keith has the power to do what he will.  But calling
> it 3.9.1 implies a blessing for his actions that he
> has not formally asked for nor formally received. Some
>  
I have never called it 3.9.1 without adding a caveat "proposed" ,
"experimental", or "unofficial"
> controls should be put on what he calls it less his
> efforts be taken (by those outside) as authorized and
> encourgaged by the offical board.
>
> Or he should be given that authorization by the board
> to go ahead.  Right now the situation is like having a
> loose canon aboard a ship.  And people should look at
> it in that manner.
>  
I find that insulting actually.

The fact is that I dont have a project I am working on or any hidden
agenda. I am aiming to serve the squeak community by developing some
stuff that is needed.

In general I am an initiating, ideas sort of person, and I value and
need support from finishing, relentless checking type of people such as
yourself.

It would be helpful if those finishing relentless checking, dotting i's
and crossing t's sort of people contributed to helping me where I am
weak, rather than shooting me down.

best regards

Keith

p.s. I hate politics

El 5/6/07 8:18 PM, "Simon Kirk" <[hidden email]> escribió:

> By "available in an image" I mean an image that a user can download with
> the changes already loaded. From the impression I had gathered before, this
> didn't seem to be easily available.

Go to http://ftp.squeak.org/3.10alpha/
Four complete and ready to use images are there.

I redo the complete process from 3.9 each time.
With a quick dirty hack I could produce MonticelloToCs (to Jerome request)
for having some similar to old style .cs from repository load.

> I'm a little worried though that I was so ignorant to what was going on
> with 3.10, but I'm ready to admit that was down to my not having read the
> list/website/weekly squeak closely enough to know. Still, I do read the
> list almost every day, so I hope my lack of knowledge isn't an example of
> others'. Let's hope it was just my blindness :)

It's not your fault.
I send many mails about 3.10 to this list , try to help in newbies list (I
wish this was a newbie ready release)
What I can't do is send [ANN] here or do things not in the plan.

Edgar

Hi keith

I'm terribly busy until the end of the week just scanning emails.
I would prefer to let 3.10 roll.

Stef

On 7 mai 07, at 03:34, Keith Hodges wrote:

> Actually I have been talking with Stephane all along about all  
> sorts of aspects of this process. Stephane has held back, in order  
> to let 3.10 roll. But if he would like to take the opportunity to  
> help put some icing on his 3.9 cake then why not.