Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » General (non-technical) » Eclipse Foundation » quality and stability of Europa
quality and stability of Europa [message #47088] Sun, 22 July 2007 19:29 Go to next message
Eclipse User
Originally posted by: bimjunk.basistech.com

Dear Eclipse Foundation,

I am an Eclipse user of relatively long standing. I have used Eclipse for a
moderately large amount of Java development. I have build plugins. I have
supplied a few patches.

And I am extremely frustrated in my experiences in trying to use any of the
layered products that have moved from the vague cloud to Callisto to Europa.

Most concretely, I rarely get more than 10 minutes into an experiment with
one of these without receiving a null pointer exception, or a blank
preference page complaining that 'the existing values are not valid' and
offering no repair mechanism. Bugzilla's submitted for these results
generally sit NEW and unresponded indefinitely. The two recent examples
here for me are WTP and STP.

The contrast between the high level of stability of the core JDT and these
items is really drastic. In my opinion, you are not doing the overall
platform any favors by positioning these immature, crash-prone projects as
if they meet the same standards as the core platform.

In addition, I will offer the opinion that there is a lack of design
coordination between these items and the core platform. Example in point:

In WTP, the most natural thing in the world is to want to use a plain,
vanilla, Java project, as a dependency of a Dynamic Web Project.

Simply listing the dependency on the normal classpath, which would be the
'least surprise' approach, is accepted with no warning but has no effect.

Instead, the POJO project must be added to a special list of 'J2ee module
dependencies'. Note, here, a problem with terminology consistency: the main
Europa page is careful to avoid the use of J2EE when the functionality
includes mere servlet containers, but WTP does not play along.

Then, it turns out that this special list of dependencies refuses to simply
add a plain Java project. Instead, it insists on adding a 'Utility project
facet' to the project. It doesn't ask for permission first. It just does so.
There is no UI to explicitly add this facet, because the facet system is
only enabled for WTP projects. There is no explanation of the side-effects
of this facet on the continued use of the POJO as such.

If WTP cannot add the facet (because, for example, the .project file is
read-only), it fails and corrupts the metadata of the WTP project.
Unrecoverably. There's not much to do except delete and recreate it.

Do you have a design review process that evaluates the consistency of these
projects with the core platform and some overall standards? It doesn't seem
so from here.

Sincerely,

Benson Margulies
CTO Basis Technology Corp.
Re: quality and stability of Europa [message #47119 is a reply to message #47088] Sun, 22 July 2007 20:09 Go to previous messageGo to next message
Eclipse User
Originally posted by: bimjunk.basistech.com

While I'm being a crank, I would also beg your indulgence in having a look
at the following bugzilla:

https://bugs.eclipse.org/bugs/show_bug.cgi?id=183601

The meat of the matter is the following: The official update stream of
eclipse 3.2.2 cannot install a signed update while running under Java 1.5.
The BIRT people have released a version that is officially set up to install
against the 3.2.2 stream that is signed. Installing it while running under
Java 1.5 fails.

The response to the bug is to instruct us to either (a) run an insecure,
obsolete, JVM, or (b) update to a prerelease of the next eclipse SDK.

Is this really acceptable under the foundation's policies? And, if not,
where other than here would one complain?
Re: quality and stability of Europa [message #47150 is a reply to message #47088] Mon, 23 July 2007 07:33 Go to previous messageGo to next message
Eclipse User
Originally posted by: merks.ca.ibm.com

Benson,

My personal comments are below. They're my own opinions and not
necessarily those of the foundation.


Benson Margulies wrote:
> Dear Eclipse Foundation,
>
> I am an Eclipse user of relatively long standing. I have used Eclipse for a
> moderately large amount of Java development. I have build plugins. I have
> supplied a few patches.
>
> And I am extremely frustrated in my experiences in trying to use any of the
> layered products that have moved from the vague cloud to Callisto to Europa.
>
You seem to be sweeping all of Callisto and Europa into the same basket
and passing judgment on all projects in them as a unit. I'm not sure
that's actually fair or balanced. Have you tried CDT for example? I
know that in frustration we're often tempted to rant about everything
that's annoying, but if you really want things to improve, you have to
be more focused on the more important specifics. It helps if you don't
tick everyone off along the way (speaking as an experienced professional
complainer).
> Most concretely, I rarely get more than 10 minutes into an experiment with
> one of these without receiving a null pointer exception, or a blank
> preference page complaining that 'the existing values are not valid' and
> offering no repair mechanism. Bugzilla's submitted for these results
> generally sit NEW and unresponded indefinitely. The two recent examples
> here for me are WTP and STP.
>
Let me assure you that EMF bugzillas don't sit in NEW state without a
response for more than a few hours nor do they typically go unfixed for
more than a week so I hope you aren't trying to generalize your
experience with particular projects as applying to all projects in
general. (You'll note how easy it is to make people feel defensive and
how little good comes from that.)
> The contrast between the high level of stability of the core JDT and these
> items is really drastic. In my opinion, you are not doing the overall
> platform any favors by positioning these immature, crash-prone projects as
> if they meet the same standards as the core platform.
>
Well, WTP has been around for a number of years and is shipped in a
number of IBM products so while I don't dispute your concerns about
quality, I would be hard to attribute that to immaturity. It's also
important to keep in mind that Europa makes no claims that any of the
projects are mature nor does it position them as meeting rigorous
standards of quality. Europa is simply the end result of a set of
independently managed projects coordinating their releases according to
a common schedule so that consumers have a more cohesive set of projects
from which to choose. There are effectively no quality standards beyond
those put in place by the individual projects themselves.
> In addition, I will offer the opinion that there is a lack of design
> coordination between these items and the core platform. Example in point:
>
> In WTP, the most natural thing in the world is to want to use a plain,
> vanilla, Java project, as a dependency of a Dynamic Web Project.
>
> Simply listing the dependency on the normal classpath, which would be the
> 'least surprise' approach, is accepted with no warning but has no effect.
>
You've opened a bugzilla about this? I opened a similar bugzilla
against the platform itself because I've noticed that a very common new
user mistake when developing plugins is to use the project properties to
set the classpath only to have the identical problem show up again at
runtime because the MANIFEST.MF should have been used to change the
dependencies instead of setting the classpath directly; this issue tends
to come up several times a week on the newsgroups and has yet to be
addressed. I agree that there is little design coordination between any
of the projects and I doubt that simply complaining about that will do
any good. Europa is only a first step towards better coordination
between projects. The foundation is taking steps to try to get the
Architecture Council to actually be a functional group that effects
change, which has not been the case to date. But I don't see any easy
solutions to this problem...
> Instead, the POJO project must be added to a special list of 'J2ee module
> dependencies'. Note, here, a problem with terminology consistency: the main
> Europa page is careful to avoid the use of J2EE when the functionality
> includes mere servlet containers, but WTP does not play along.
>
Complaining to the foundation about WTP's design problems is highly
unlikely to result the kind of change you are looking for. Perhaps
discussions with the developers have proven futile, but that's really
the only way to effect change. Getting directly involved would be the
best way. It's not as if the developers are employees of the foundation
and the foundation can direct their activities as it sees fit. And
there may well be reasons why WTP needs additional metadata in the
projects in order for them to work properly at runtime, but maybe not.
> Then, it turns out that this special list of dependencies refuses to simply
> add a plain Java project. Instead, it insists on adding a 'Utility project
> facet' to the project. It doesn't ask for permission first. It just does so.
> There is no UI to explicitly add this facet, because the facet system is
> only enabled for WTP projects. There is no explanation of the side-effects
> of this facet on the continued use of the POJO as such.
>
I believe WTP was seeking to have this facet design push down into the
platform. Would that be a good idea or a bad one in your opinion? If a
bad one, then keep in mind that coordination will not necessarily
produce a good design.
> If WTP cannot add the facet (because, for example, the .project file is
> read-only), it fails and corrupts the metadata of the WTP project.
> Unrecoverably. There's not much to do except delete and recreate it.
>
That sounds bad. You've opened a bugzilla for that? I know it might
sit around for a long time, but if it's not in the queue, folks might
not even know there is a problem like this. With a projects like EMF
and JDT, I believe the overall quality of the project benefits because
we are forced to use the project ourselves to do our own development. I
think many other projects have more problems because the developers
don't use the project to do their development so the project isn't
really in continuous testing mode as is the case for projects that
bootstrap their own development. Most projects are badly understaffed
and quality almost always suffers in that context.
> Do you have a design review process that evaluates the consistency of these
> projects with the core platform and some overall standards? It doesn't seem
> so from here.
>
The problem to me doesn't sound like merely a problem of consistency
with the core platform, which in the case of modifying the classpath of
a plugin project allows users to fall into a similarly confusing trap.
There is a release review process that's open to the community and this
is a potential place to air grievances such as the one you've expressed,
but no one shows up for them. Most PMC's meet monthly, so you might
want to ask to be included in one of those meetings to discuss specific
issues you have with that project. As an elected committer
representative on Eclipse's board of directors, I share many of the
concerns you've expressed and I'm always concerned when the quality of
any one Eclipse project affects the overall perception of Eclipse being
a place that produces quality results. I'm just not sure what the
solution to the problem would be. It's certainly not as simple as
having reviews when many projects already have longs lists of known and
outstanding problems that already can't be addressed because of resource
constraints. The only real solution seems to be more committers with
the time, energy, and motivation to make improvements.
> Sincerely,
>
> Benson Margulies
> CTO Basis Technology Corp.
>
>
>
Re: quality and stability of Europa [message #47179 is a reply to message #47119] Mon, 23 July 2007 07:52 Go to previous messageGo to next message
Eclipse User
Originally posted by: merks.ca.ibm.com

Benson,

If I understand this problem properly, it's not one that BIRT can fix
since none of their code is in the stack trace. BIRT along with every
other Europa project was mandated to sign jars (a rule that was later
relaxed because everyone couldn't meet the requirement) and if older
versions of the update manager on a 5.0 JVM can't handle that, then
that's very unfortunate but beyond BIRT's direct control. What exactly
did you expect from them directly? (I don't suppose the one day
response time made you happy in this case.) If BIRT is supposed to run
on older versions of Eclipse, then you perhaps could unzip the plugins
rather than use the update manager. Given that you were trying to get
a prerelease of BIRT at the time, using a prerelease of Eclipse for it
doesn't strike me as at all unreasonable.

You seem to be under the impression that the foundation can make
projects do as the foundation directs but in reality, the committers can
do pretty much as they please (within reason) as long as they follow all
the appropriate intellectual property guidelines. So if you think you
can best accomplish your goals by getting the foundation to beat
projects with a big stick it's good to keep in mind that the foundation
doesn't have a big stick in this regard. In the open source world, most
things are accomplished by asking nicely and by offering to do things in
exchange for the things you get...


Benson Margulies wrote:
> While I'm being a crank, I would also beg your indulgence in having a look
> at the following bugzilla:
>
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=183601
>
> The meat of the matter is the following: The official update stream of
> eclipse 3.2.2 cannot install a signed update while running under Java 1.5.
> The BIRT people have released a version that is officially set up to install
> against the 3.2.2 stream that is signed. Installing it while running under
> Java 1.5 fails.
>
> The response to the bug is to instruct us to either (a) run an insecure,
> obsolete, JVM, or (b) update to a prerelease of the next eclipse SDK.
>
> Is this really acceptable under the foundation's policies? And, if not,
> where other than here would one complain?
>
>
>
Re: quality and stability of Europa [message #47194 is a reply to message #47150] Mon, 23 July 2007 23:33 Go to previous messageGo to next message
Eclipse User
Originally posted by: bimjunk.basistech.com

Ed,

I descended into this minor bit of grandstanding because of the hype
surrounding Europa. I apologize for the accidental tarring of projects that
I don't know (or care) much about, but I was trying to make a point about
the foundations's policies for the overall package using the particular
cases I've encountered. I'll elaborate on my point of view. On the signed
jars problem, I'm (unrhetorically) confused by the information in your
response. I'll explain at the end.

So, follow me in a trip back down memory lane.

Pre-Callisto, there were these layered products. Under the Elipse umbrella,
sure, but clearly independent. If one of them didn't work very well, that
wasn't offensive. They were all (a) new, (b) managed under separate cover.

Fast forward to Callisto. Callisto is was the first time that the Eclipse
web site proposed to offer this bundle of disparate technologies as being,
collectively, somewhat ready for prime time, interoperable, and all of that.

As you may recall, anyone who did what I did, and installed all of Callisto
from the update site, got a very rude surprise: a set of sample plugins
(samples in the sense of illustrating plugin development) were included in
the Callisto update site, and resulted in ugly and mysterious green
decorations. I recall adding my voice to a small avalanche of bugzillas
submitted in that incident, and learning the lesson that the Callisto update
site had to be approached with some care.

Now we reach Europa. Europa is rolled out with a much louder collection of
hype, including a lot of punditry about the war between Eclipse and
netbeans. All of the sudden, the main Eclipse download site proposes that
the Europa bundles have the same status of endorsement as the core SDK. As I
complained in a bz, the whole Eclipse version number of '3.3' is nearly
completely absent from that page. According to the official first-responder
to that bugzilla, the numbered version scheme is more or less passe, and
Europa is now the official main Eclipse distribution.

To me, that engenders an expectation: that the Europa downloads offered on
the main download page will deliver the same level of stability and
coherence as the Eclipse 3.2 download packages did.

So, when I download and install a Europa bundle, and discover that WTP (a)
crashes just like the version last seen on Callisto, and (b) seems to my
jaundiced and perhaps uninformed view to be in conflict with the core
platform on classpath management, I fulminated for a few days. Then I
decided to post to this newsgroup.

I hear you loud and clear that the foundation, like any open source
operation, has limited control over the activities of the various projects.
But the foundation has 100% control over the decision to agglomerate a set
of projects into Europa and treat Europa as the main release of Eclipse,
instead of repeating the Callisto pattern of an off-to-one-side repository
of additional function.

At the bottom line, I offer this information to the readers of this message,
to do with as they wish: the marketing of Europa has made Eclipse as a
trademark or brand name seem less dependable to me. I might be unique. I
might not.

As for jar signing: my complaint is not (now) with the BIRT people. It is
with the larger policy that you cite, which the BIRT people did not point us
sufferers to. Here we have a release of Eclipse (3.2) that is (as far as I
know) officially supported on Java 1.5. That would seem to me to require
patches to this release to be released compatibly with 1.5. As far as I
know, each of the BIRT-like projects has multiple release streams
corresponding to the Eclipse main versions: one for 3.2, one for 3.3. Why
does the policy call for them to start emitting these problematic signed
packages into the 3.2 stream if someone isn't going to first patch 3.2 to be
able to install them? I know that I'm not the only user frustrated by this
state of affairs, since I've gotten email messages from fellow-suffers
wondering if I've gotten any relief.

That's my story. I don't see this as an argument, let alone as an argument
that I could possibly win. If this is nourishing food for thought, that's
good. If it plays as cranky flaming, then I've done a bad job of presenting
my thoughts. If there's a more constructive forum for me to politely weigh
in on questions like the role of facets, lead me to it.

--benson margulies
Re: quality and stability of Europa [message #47218 is a reply to message #47194] Tue, 24 July 2007 07:00 Go to previous messageGo to next message
Eclipse User
Originally posted by: merks.ca.ibm.com

Benson,

This story sounds far more balanced and constructive than your previous
posts. I think you've given folks some important things to think about
and to act upon. Many of us work hard to achieve quality and we would
like to see Eclipse as a brand that is associated with quality. I'm
sure even in cases where quality is a little (or perhaps even a lot)
below par that folks are working hard in a resource constrained
environment to achieve the best results possible. I agree with you that
when things are offered as a packaged bundle, users should expect the
things in that package to be effectively certified as being of the same
high quality that folks have grown to love and expect from the base
platform bundle they've been downloading for years. I certainly
intended to follow up on this issue with the other committer
representatives on the Eclipse board and if possible with the board
itself. I'm also a member of the Planning and Architecture councils,
both of which need to take this issue very seriously. The Planning
council in particular will need to focus beyond just achieving
deadlines, which we've quite good at now, and look toward usability,
integration, and of course overall quality. I'm not sure what the best
solution would be because I'm not sure the extent of the quality problem
nor even the extent to which quality problems are a result of lack of
testing on all platforms or the result of lack of resource to fix the
known problems.

With respect to jar signing, I wouldn't expect folks to begin signing of
maintenance stream builds if loading the resulting jars in a know
problem in the target platform of those builds. Certainly no one was
required to do signing for anything other than their current stream
targeted for Eclipse 3.3, and in the end, that requirement was dropped
completely though I expect it will be an absolute requirement for Ganymede.

Thanks for taking the time to recompose your thoughts...


Benson Margulies wrote:
> Ed,
>
> I descended into this minor bit of grandstanding because of the hype
> surrounding Europa. I apologize for the accidental tarring of projects that
> I don't know (or care) much about, but I was trying to make a point about
> the foundations's policies for the overall package using the particular
> cases I've encountered. I'll elaborate on my point of view. On the signed
> jars problem, I'm (unrhetorically) confused by the information in your
> response. I'll explain at the end.
>
> So, follow me in a trip back down memory lane.
>
> Pre-Callisto, there were these layered products. Under the Elipse umbrella,
> sure, but clearly independent. If one of them didn't work very well, that
> wasn't offensive. They were all (a) new, (b) managed under separate cover.
>
> Fast forward to Callisto. Callisto is was the first time that the Eclipse
> web site proposed to offer this bundle of disparate technologies as being,
> collectively, somewhat ready for prime time, interoperable, and all of that.
>
> As you may recall, anyone who did what I did, and installed all of Callisto
> from the update site, got a very rude surprise: a set of sample plugins
> (samples in the sense of illustrating plugin development) were included in
> the Callisto update site, and resulted in ugly and mysterious green
> decorations. I recall adding my voice to a small avalanche of bugzillas
> submitted in that incident, and learning the lesson that the Callisto update
> site had to be approached with some care.
>
> Now we reach Europa. Europa is rolled out with a much louder collection of
> hype, including a lot of punditry about the war between Eclipse and
> netbeans. All of the sudden, the main Eclipse download site proposes that
> the Europa bundles have the same status of endorsement as the core SDK. As I
> complained in a bz, the whole Eclipse version number of '3.3' is nearly
> completely absent from that page. According to the official first-responder
> to that bugzilla, the numbered version scheme is more or less passe, and
> Europa is now the official main Eclipse distribution.
>
> To me, that engenders an expectation: that the Europa downloads offered on
> the main download page will deliver the same level of stability and
> coherence as the Eclipse 3.2 download packages did.
>
> So, when I download and install a Europa bundle, and discover that WTP (a)
> crashes just like the version last seen on Callisto, and (b) seems to my
> jaundiced and perhaps uninformed view to be in conflict with the core
> platform on classpath management, I fulminated for a few days. Then I
> decided to post to this newsgroup.
>
> I hear you loud and clear that the foundation, like any open source
> operation, has limited control over the activities of the various projects.
> But the foundation has 100% control over the decision to agglomerate a set
> of projects into Europa and treat Europa as the main release of Eclipse,
> instead of repeating the Callisto pattern of an off-to-one-side repository
> of additional function.
>
> At the bottom line, I offer this information to the readers of this message,
> to do with as they wish: the marketing of Europa has made Eclipse as a
> trademark or brand name seem less dependable to me. I might be unique. I
> might not.
>
> As for jar signing: my complaint is not (now) with the BIRT people. It is
> with the larger policy that you cite, which the BIRT people did not point us
> sufferers to. Here we have a release of Eclipse (3.2) that is (as far as I
> know) officially supported on Java 1.5. That would seem to me to require
> patches to this release to be released compatibly with 1.5. As far as I
> know, each of the BIRT-like projects has multiple release streams
> corresponding to the Eclipse main versions: one for 3.2, one for 3.3. Why
> does the policy call for them to start emitting these problematic signed
> packages into the 3.2 stream if someone isn't going to first patch 3.2 to be
> able to install them? I know that I'm not the only user frustrated by this
> state of affairs, since I've gotten email messages from fellow-suffers
> wondering if I've gotten any relief.
>
> That's my story. I don't see this as an argument, let alone as an argument
> that I could possibly win. If this is nourishing food for thought, that's
> good. If it plays as cranky flaming, then I've done a bad job of presenting
> my thoughts. If there's a more constructive forum for me to politely weigh
> in on questions like the role of facets, lead me to it.
>
> --benson margulies
>
>
>
>
>
Re: quality and stability of Europa [message #47245 is a reply to message #47194] Tue, 24 July 2007 11:52 Go to previous messageGo to next message
Gunnar Wagenknecht is currently offline Gunnar Wagenknecht
Messages: 447
Registered: July 2009
Senior Member
Benson Margulies wrote:
> As I complained in a bz, the whole Eclipse version number of '3.3' is nearly
> completely absent from that page. According to the official first-responder
> to that bugzilla, the numbered version scheme is more or less passe, and
> Europa is now the official main Eclipse distribution.

I object. I did not use the word 'passé' nor did I purport that the
numbered versioning scheme is "more or less" gone.

https://bugs.eclipse.org/bugs/show_bug.cgi?id=196539

Anyway, there are some points raised that I do agree with. Eclipse is
not bug-free and some concept can be optimized in terms of integration,
design, etc. But even Rome wasn't built in a day and neither will one
release cycle be sufficient to solve all issues. However, my personal
feeling is that Europa is significantly better than Callisto. But that
may really just be my personal feeling based on the Eclipse
functionality I use.

-Gunnar

--
Gunnar Wagenknecht
gunnar@wagenknecht.org
http://wagenknecht.org/
Re: quality and stability of Europa [message #47286 is a reply to message #47088] Tue, 24 July 2007 12:04 Go to previous messageGo to next message
David Williams is currently offline David Williams
Messages: 696
Registered: July 2009
Senior Member
On Sun, 22 Jul 2007 19:29:23 -0400, Benson Margulies <bimjunk@basistech.com> wrote:

> Dear Eclipse Foundation,
>
[....]
>
> Sincerely,
>
> Benson Margulies
> CTO Basis Technology Corp.
>

Benson,

Thanks for your comments. Some of the problems you report in your posts about WTP do sound serious, and am surprised I haven't heard of them before. I checked bugzilla and our webtools newsgroup and couldn't find any reports from you there, or about these bugs (with my quick scan) ... so, guess you know, it would be best to open bugs on issues you find. Any WTP specific issues should be addressed via bug reports or the webtools newsgroup, but seemed your post here to the Foundation newsgroup was to question the Foundation on the meaning, purpose, and process of the yearly coordinated releases. So, I'll make a few general comments on that.

Perhaps the marketing messages mismatched reality (that is rare, though, right? :) but the primary emphasis of our yearly simultaneous releases is just that -- simultaneous release. And the theory is (or, should I say, the hope is) that by doing that, we will at least promote more "simultaneous use" which in years past was literally impossible, since various versions of various projects never did work together at all, except for some special cases at some very specific windows of time.

Many of us were disappointed that more testing wasn't done with Europa as the milestones progressed, by the projects and by the community, and that many of us projects found it necessary to be making pretty big changes near the end, instead of a gradual smooth shutdown. So, we at Eclipse are hoping to improve that next time around. However, there actually were some pretty big "cross-project" bugs fixed, that would have not been possible to fix (to detect in time, that is) if we had not done the simultaneous release ... and, for good and bad, many of those worst bugs are never seen: good since they should not be seen, bad since it's hard for others to see the progress made ... meager as it might be.

So, again just want to say thanks, we value all comments, positive and critical. The more specific comments you can make, in bugzillas, the better it will be, and the more you can participate throughout the release cycle, the better it will be. I know not everyone can afford to participate throughout a release cycle, and can just most easily make comments right after a release, but I think the success/failure of Ganymede (and Eclipse as a whole?) depends on that sort of on-going participation, so, I hope you, and other readers of this thread, can join in that sort of participation for Ganymede.
Re: quality and stability of Europa [message #47305 is a reply to message #47179] Tue, 24 July 2007 13:09 Go to previous messageGo to next message
Eclipse User
Originally posted by: alex_blewitt.yahoo.com

> BIRT along with every other Europa project was mandated
> to sign jars ...

> ... you seem to be under the impression that the foundation
> can make projects do as the foundation directs ...

Surely this is a contradiction, apparent or otherwise? Granted, the foundation can't drive projects; but who was mandating the signing of Jars in that case?

Alex.
Re: quality and stability of Europa [message #47347 is a reply to message #47305] Tue, 24 July 2007 13:41 Go to previous messageGo to next message
Eclipse User
Originally posted by: merks.ca.ibm.com

This is a multi-part message in MIME format.
--------------020700040507020805040304
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Alex,

The teams working on Europa collectively agree to the requirements that
will be expected of all the Europa projects. They are not simply
dictated to us by the foundation or anyone else. These requirements
focus on reasonable expectations that will make the overall result
better, rather than on exclusionary rules that might limit
participation. The only way to enforce these rules is to threaten to
kick folks off the train. And that's typically a hollow threat as I've
pointed out to folks. You can kick off the caboose, but you can't kick
off one of the locomotives; i.e., a release train without EMF isn't
possible. So threats are ineffective which just leaves us with the
needing to convince people to do the right things and hoping that
cooperation will prevail. That's proven to work well so far.


Alex Blewitt wrote:
>> BIRT along with every other Europa project was mandated
>> to sign jars ...
>>
>
>
>> ... you seem to be under the impression that the foundation
>> can make projects do as the foundation directs ...
>>
>
> Surely this is a contradiction, apparent or otherwise? Granted, the foundation can't drive projects; but who was mandating the signing of Jars in that case?
>
> Alex.
>


--------------020700040507020805040304
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Alex,<br>
<br>
The teams working on Europa collectively agree to the requirements that
will be expected of all the Europa projects.&nbsp;&nbsp; They are not simply
dictated to us by the foundation or anyone else.&nbsp; These requirements
focus on reasonable expectations that will make the overall result
better, rather than on exclusionary rules that might limit
participation.&nbsp; The only way to enforce these rules is to threaten to
kick folks off the train.&nbsp; And that's typically a hollow threat as I've
pointed out to folks.&nbsp; You can kick off the caboose, but you can't kick
off one of the locomotives; i.e., a release train without EMF isn't
possible.&nbsp; So threats are ineffective which just leaves us with the
needing to convince people to do the right things and hoping that
cooperation will prevail.&nbsp; That's proven to work well so far.<br>
<br>
<br>
Alex Blewitt wrote:
<blockquote
cite="mid:569920148.8811185296989127.JavaMail.root@cp9.dzone.com"
type="cite">
<blockquote type="cite">
<pre wrap="">BIRT along with every other Europa project was mandated
to sign jars ...
</pre>
</blockquote>
<pre wrap=""><!---->
</pre>
<blockquote type="cite">
<pre wrap="">... you seem to be under the impression that the foundation
can make projects do as the foundation directs ...
</pre>
</blockquote>
<pre wrap=""><!---->
Surely this is a contradiction, apparent or otherwise? Granted, the foundation can't drive projects; but who was mandating the signing of Jars in that case?

Alex.
</pre>
</blockquote>
<br>
</body>
</html>

--------------020700040507020805040304--
Re: quality and stability of Europa [message #47361 is a reply to message #47088] Tue, 24 July 2007 20:07 Go to previous messageGo to next message
Eclipse User
Originally posted by: atoulme.intalio.com

Hi,

I am a committer over the STP project, working on the BPMN modeler.
I just checked the bugzilla and did not see any bugs from you regarding
our project (I looked for your name in the open bugs only, so I might
have missed something).

If you experienced any trouble, we will be able to assist you, just drop
by the newsgroup or file a bug.

Yours,

Antoine Toulme

Benson Margulies wrote:
> Dear Eclipse Foundation,
>
> I am an Eclipse user of relatively long standing. I have used Eclipse for a
> moderately large amount of Java development. I have build plugins. I have
> supplied a few patches.
>
> And I am extremely frustrated in my experiences in trying to use any of the
> layered products that have moved from the vague cloud to Callisto to Europa.
>
> Most concretely, I rarely get more than 10 minutes into an experiment with
> one of these without receiving a null pointer exception, or a blank
> preference page complaining that 'the existing values are not valid' and
> offering no repair mechanism. Bugzilla's submitted for these results
> generally sit NEW and unresponded indefinitely. The two recent examples
> here for me are WTP and STP.
>
> The contrast between the high level of stability of the core JDT and these
> items is really drastic. In my opinion, you are not doing the overall
> platform any favors by positioning these immature, crash-prone projects as
> if they meet the same standards as the core platform.
>
> In addition, I will offer the opinion that there is a lack of design
> coordination between these items and the core platform. Example in point:
>
> In WTP, the most natural thing in the world is to want to use a plain,
> vanilla, Java project, as a dependency of a Dynamic Web Project.
>
> Simply listing the dependency on the normal classpath, which would be the
> 'least surprise' approach, is accepted with no warning but has no effect.
>
> Instead, the POJO project must be added to a special list of 'J2ee module
> dependencies'. Note, here, a problem with terminology consistency: the main
> Europa page is careful to avoid the use of J2EE when the functionality
> includes mere servlet containers, but WTP does not play along.
>
> Then, it turns out that this special list of dependencies refuses to simply
> add a plain Java project. Instead, it insists on adding a 'Utility project
> facet' to the project. It doesn't ask for permission first. It just does so.
> There is no UI to explicitly add this facet, because the facet system is
> only enabled for WTP projects. There is no explanation of the side-effects
> of this facet on the continued use of the POJO as such.
>
> If WTP cannot add the facet (because, for example, the .project file is
> read-only), it fails and corrupts the metadata of the WTP project.
> Unrecoverably. There's not much to do except delete and recreate it.
>
> Do you have a design review process that evaluates the consistency of these
> projects with the core platform and some overall standards? It doesn't seem
> so from here.
>
> Sincerely,
>
> Benson Margulies
> CTO Basis Technology Corp.
>
>


--
Intalio, the Open Source BPMS Company

<a href="http://www.intalio.com">http://www.intalio.com</a>
<a href="http://bpms.intalio.com">Community website</a>
Re: quality and stability of Europa [message #47389 is a reply to message #47286] Wed, 25 July 2007 06:42 Go to previous messageGo to next message
Eclipse User
Originally posted by: merks.ca.ibm.com

This is a multi-part message in MIME format.
--------------020002050505030709070906
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

David,

Probably you didn't find any because Benson is using a slightly
different ID than he has for these newsgroup postings. The following
query lists all the bugs Benson or anyone from basistech.com has
reported for all Eclipse projects:

https://bugs.eclipse.org/bugs/buglist.cgi?query_format=advan ced&short_desc_type=allwordssubstr&short_desc=&l ong_desc_type=allwordssubstr&long_desc=&bug_file_loc _type=allwordssubstr&bug_file_loc=&status_whiteboard _type=allwordssubstr&status_whiteboard=&keywords_typ e=allwords&keywords=&emailreporter1=1&emailtype1 =substring&email1=basistech.com&emailtype2=substring &email2=&bugidtype=include&bug_id=&votes=&am p;chfieldfrom=&chfieldto=Now&chfieldvalue=&cmdty pe=doit&order=Reuse+same+sort+as+last+time&field0-0- 0=noop&type0-0-0=noop&value0-0-0=
< https://bugs.eclipse.org/bugs/buglist.cgi?query_format=advan ced&short_desc_type=allwordssubstr&short_desc=&l ong_desc_type=allwordssubstr&long_desc=&bug_file_loc _type=allwordssubstr&bug_file_loc=&status_whiteboard _type=allwordssubstr&status_whiteboard=&keywords_typ e=allwords&keywords=&emailreporter1=1&emailtype1 =substring&email1=basistech.com&emailtype2=substring &email2=&bugidtype=include&bug_id=&votes=&am p;chfieldfrom=&chfieldto=Now&chfieldvalue=&cmdty pe=doit&order=Reuse+same+sort+as+last+time&field0-0- 0=noop&type0-0-0=noop&value0-0-0=>

There are a few Web Tools bugzillas, but not many, and certainly not
many old ones that appear to be serious problems that have have simply
been ignored...

For little humor, it's interesting to see how the User Assistance
category is truncated to just after the second "s" of the second word in
the tabular view. :-P


David Williams wrote:
> On Sun, 22 Jul 2007 19:29:23 -0400, Benson Margulies <bimjunk@basistech.com> wrote:
>
>
>> Dear Eclipse Foundation,
>>
>>
> [....]
>
>> Sincerely,
>>
>> Benson Margulies
>> CTO Basis Technology Corp.
>>
>>
>
> Benson,
>
> Thanks for your comments. Some of the problems you report in your posts about WTP do sound serious, and am surprised I haven't heard of them before. I checked bugzilla and our webtools newsgroup and couldn't find any reports from you there, or about these bugs (with my quick scan) ... so, guess you know, it would be best to open bugs on issues you find. Any WTP specific issues should be addressed via bug reports or the webtools newsgroup, but seemed your post here to the Foundation newsgroup was to question the Foundation on the meaning, purpose, and process of the yearly coordinated releases. So, I'll make a few general comments on that.
>
> Perhaps the marketing messages mismatched reality (that is rare, though, right? :) but the primary emphasis of our yearly simultaneous releases is just that -- simultaneous release. And the theory is (or, should I say, the hope is) that by doing that, we will at least promote more "simultaneous use" which in years past was literally impossible, since various versions of various projects never did work together at all, except for some special cases at some very specific windows of time.
>
> Many of us were disappointed that more testing wasn't done with Europa as the milestones progressed, by the projects and by the community, and that many of us projects found it necessary to be making pretty big changes near the end, instead of a gradual smooth shutdown. So, we at Eclipse are hoping to improve that next time around. However, there actually were some pretty big "cross-project" bugs fixed, that would have not been possible to fix (to detect in time, that is) if we had not done the simultaneous release ... and, for good and bad, many of those worst bugs are never seen: good since they should not be seen, bad since it's hard for others to see the progress made ... meager as it might be.
>
> So, again just want to say thanks, we value all comments, positive and critical. The more specific comments you can make, in bugzillas, the better it will be, and the more you can participate throughout the release cycle, the better it will be. I know not everyone can afford to participate throughout a release cycle, and can just most easily make comments right after a release, but I think the success/failure of Ganymede (and Eclipse as a whole?) depends on that sort of on-going participation, so, I hope you, and other readers of this thread, can join in that sort of participation for Ganymede.
>
>


--------------020002050505030709070906
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
David,<br>
<br>
Probably you didn't find any because Benson is using a slightly
different ID than he has for these newsgroup postings.&nbsp; The following
query lists all the bugs Benson or anyone from basistech.com has
reported for all Eclipse projects:<br>
<blockquote><a
href=" https://bugs.eclipse.org/bugs/buglist.cgi?query_format=advan ced&amp ;short_desc_type=allwordssubstr&amp;short_desc=&amp; long_desc_type=allwordssubstr&amp;long_desc=&amp;bug _file_loc_type=allwordssubstr&amp;bug_file_loc=&amp; status_whiteboard_type=allwordssubstr&amp;status_whitebo ard=&amp;keywords_type=allwords&amp;keywords=&am p;emailreporter1=1&amp;emailtype1=substring&amp;emai l1=basistech.com&amp;emailtype2=substring&amp;email2 =&amp;bugidtype=include&amp;bug_id=&amp;votes=&a mp;amp;chfieldfrom=&amp;chfieldto=Now&amp;chfieldval ue=&amp;cmdtype=doit&amp;order=Reuse+same+sort+as+la st+time&amp;field0-0-0=noop&amp;type0-0-0=noop&a mp;value0-0-0= "> https://bugs.eclipse.org/bugs/buglist.cgi?query_format=advan ced&amp ;short_desc_type=allwordssubstr&amp;short_desc=&amp; long_desc_type=allwordssubstr&amp;long_desc=&amp;bug _file_loc_type=allwordssubstr&amp;bug_file_loc=&amp; status_whiteboard_type=allwordssubstr&amp;status_whitebo ard=&amp;keywords_type=allwords&amp;keywords=&am p;emailrepor
ter1=1&amp;emailtype1=substring&amp;email1=basistech .com&amp;emailtype2=substring&amp;email2=&amp;bu gidtype=include&amp;bug_id=&amp;votes=&amp;chfie ldfrom=&amp;chfieldto=Now&amp;chfieldvalue=&amp; cmdtype=doit&amp;order=Reuse+same+sort+as+last+time& amp;field0-0-0=noop&amp;type0-0-0=noop&amp;value0-0- 0= </a><br>
</blockquote>
There are a few Web Tools bugzillas, but not many, and certainly not
many old ones that appear to be serious problems that have have simply
been ignored...<br>
<br>
For little humor, it's interesting to see how the User Assistance
category is truncated to just after the second "s" of the second word
in the tabular view.&nbsp; :-P<br>
<br>
<br>
David Williams wrote:
<blockquote cite="mid:op.tvy790toac05ss@dmw2t23.ibm.com" type="cite">
<pre wrap="">On Sun, 22 Jul 2007 19:29:23 -0400, Benson Margulies <a class="moz-txt-link-rfc2396E" href="mailto:bimjunk@basistech.com">&lt;bimjunk@basistech.com&gt;</a> wrote:

</pre>
<blockquote type="cite">
<pre wrap="">Dear Eclipse Foundation,

</pre>
</blockquote>
<pre wrap=""><!---->[....]
</pre>
<blockquote type="cite">
<pre wrap="">Sincerely,

Benson Margulies
CTO Basis Technology Corp.

</pre>
</blockquote>
<pre wrap=""><!---->
Benson,

Thanks for your comments. Some of the problems you report in your posts about WTP do sound serious, and am surprised I haven't heard of them before. I checked bugzilla and our webtools newsgroup and couldn't find any reports from you there, or about these bugs (with my quick scan) ... so, guess you know, it would be best to open bugs on issues you find. Any WTP specific issues should be addressed via bug reports or the webtools newsgroup, but seemed your post here to the Foundation newsgroup was to question the Foundation on the meaning, purpose, and process of the yearly coordinated releases. So, I'll make a few general comments on that.

Perhaps the marketing messages mismatched reality (that is rare, though, right? :) but the primary emphasis of our yearly simultaneous releases is just that -- simultaneous release. And the theory is (or, should I say, the hope is) that by doing that, we will at least promote more "simultaneous use" which in years past was literally impossible, since various versions of various projects never did work together at all, except for some special cases at some very specific windows of time.

Many of us were disappointed that more testing wasn't done with Europa as the milestones progressed, by the projects and by the community, and that many of us projects found it necessary to be making pretty big changes near the end, instead of a gradual smooth shutdown. So, we at Eclipse are hoping to improve that next time around. However, there actually were some pretty big "cross-project" bugs fixed, that would have not been possible to fix (to detect in time, that is) if we had not done the simultaneous release ... and, for good and bad, many of those worst bugs are never seen: good since they should not be seen, bad since it's hard for others to see the progress made ... meager as it might be.

So, again just want to say thanks, we value all comments, positive and critical. The more specific comments you can make, in bugzillas, the better it will be, and the more you can participate throughout the release cycle, the better it will be. I know not everyone can afford to participate throughout a release cycle, and can just most easily make comments right after a release, but I think the success/failure of Ganymede (and Eclipse as a whole?) depends on that sort of on-going participation, so, I hope you, and other readers of this thread, can join in that sort of participation for Ganymede.

</pre>
</blockquote>
<br>
</body>
</html>

--------------020002050505030709070906--
Re: quality and stability of Europa [message #47419 is a reply to message #47389] Wed, 25 July 2007 10:32 Go to previous messageGo to next message
Eclipse User
Originally posted by: bimjunk.basistech.com

This is a multi-part message in MIME format.

------=_NextPart_000_0016_01C7CEA7.2A931110
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

You won't find many bugs for me on WTP, because the first one stopped me =
in my tracks.

https://bugs.eclipse.org/bugs/show_bug.cgi?id=3D197222 is the crash, and =
197221 registers my beef about the use of facets.

You might ask, where are all the bugs from the past? Well, honestly, in =
the past I could never convince myself that I was going to add anything =
by complaining about issues which, at the time, just seemed to be things =
that the development team hadn't gotten around to yet.=20

For STP, there's 197426, which made it kind of difficult to install the =
thing at all. (Not an STP problem, a GMF problem). Why nothing after =
that? After my initial all-too-trollish posting here, I realized that =
STP was just hitting the same NPE in WTP (wheeee....) that i complained =
about in the bugzilla mentioned above.

As for the question of 196539, all I can say is that when I read the =
response, the impression I got what that the plan was to make Europa the =
lead version label for users of Eclipse, as opposed to people who =
develop Eclipse or plugins. If I misinterpreted the intent of your =
comments, I'm sorry. The net effect of the download page, in my entirely =
personal opinion, is consistent with remark that you found =
objectionable, but it's not really a question that deserves a debate.

I will add, before wandering back into the sand, that I posted my =
original message after (and I do not exagerate) several hours of trying =
to set up WTP without having it get into the broken state described in =
197222. That might not justify my tone, but perhaps you might have a =
momentary feeling of sympathy.

------=_NextPart_000_0016_01C7CEA7.2A931110
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type =
content=3Dtext/html;charset=3DISO-8859-1>
<META content=3D"MSHTML 6.00.6000.16481" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY text=3D#000000 bgColor=3D#ffffff>
<DIV>You won't find many bugs for me on WTP, because the first one =
stopped me in=20
my tracks.</DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=3DArial size=3D2><A=20
href=3D"https://bugs.eclipse.org/bugs/show_bug.cgi?id=3D197222">https://b=
ugs.eclipse.org/bugs/show_bug.cgi?id=3D197222</A>&nbsp;is=20
the crash, and 197221 registers my beef about the use of =
facets.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DArial size=3D2>You might ask, where are all the bugs =
from the=20
past? Well, honestly, in the past I could never convince myself that I =
was going=20
to add anything by complaining about issues which, at the time, just =
seemed to=20
be things that the development team hadn't gotten around to yet. =
</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT><FONT face=3DArial =
size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DArial size=3D2>For STP, there's 197426, which made it =
kind of=20
difficult to install the thing at all. (Not an STP problem, a GMF=20
problem).&nbsp;Why nothing after that? After my initial all-too-trollish =
posting=20
here, I realized that STP was just hitting the same NPE in WTP =
(wheeee....) that=20
i complained about in the bugzilla mentioned above.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DArial size=3D2>As for the question of 196539, all I =
can say is=20
that when I read the response, the impression I got what that the plan =
was to=20
make Europa the lead version label for users of Eclipse, as opposed to =
people=20
who develop Eclipse or plugins. If I misinterpreted the intent of your =
comments,=20
I'm sorry. The net effect of the download page, in my entirely personal =
opinion,=20
is consistent with remark that you found objectionable, but it's not =
really a=20
question that deserves a debate.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DArial size=3D2>I will add, before wandering back into =
the sand,=20
that I posted my original message after (and I do not exagerate) several =
hours=20
of trying to set up WTP without having it get into the broken state =
described in=20
197222. That might not justify my tone, but perhaps you might have a =
momentary=20
feeling of sympathy.</FONT></DIV>
<DIV>&nbsp;</DIV></BODY></HTML>

------=_NextPart_000_0016_01C7CEA7.2A931110--
Re: quality and stability of Europa [message #47449 is a reply to message #47419] Wed, 25 July 2007 11:52 Go to previous messageGo to next message
Eclipse User
Originally posted by: atoulme.intalio.com

First, thanks Ed for the query.

I could not find any open bugs against STP in the list though.

> For STP, there's 197426, which made it kind of difficult to install the
> thing at all. (Not an STP problem, a GMF problem).

This problem is due to a page rendering (that I have not been able to
reproduce so far) and does not prevent you from installing GMF at all.

All you need is to go to the update manager in Eclipse, then select the
STP features you are interested in, and click on Select Required.

You should not have to worry about anything else.
Why nothing after
> that? After my initial all-too-trollish posting here, I realized that
> STP was just hitting the same NPE in WTP (wheeee....) that i complained
> about in the bugzilla mentioned above.
You can install STP without installing WTP, at least for the BPMN modeler.

I hope this helps you.

Antoine
--
Intalio, the Open Source BPMS Company

<a href="http://www.intalio.com">http://www.intalio.com</a>
<a href="http://bpms.intalio.com">Community website</a>
Re: quality and stability of Europa [message #47479 is a reply to message #47419] Wed, 25 July 2007 17:23 Go to previous messageGo to next message
David Williams is currently offline David Williams
Messages: 696
Registered: July 2009
Senior Member
On Wed, 25 Jul 2007 10:32:57 -0400, Benson Margulies <bimjunk@basistech.=
com> wrote:

> You won't find many bugs for me on WTP, because the first one stopped =
me in my tracks.
>
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=3D197222 is the crash, a=
nd 197221 registers my beef about the use of facets.
>
[...]
> but perhaps you might have a momentary feeling of sympathy.
>

yes, you have my sympathy, and I will be extremely embarrased if we miss=
ed this simplest testing of "pessimisic" source code management systems!=


The bug links do help me understand better what you meant, and I'll foll=
ow up on those. But if we have trouble reproducing, we'll need more deta=
ils from you there.

Meanwhile, if you (or other readers of this general foundation thread) a=
re interested in diving deeper,
while we don't have a "cross project" newsgroup, there is a "cross proje=
ct" mailing list
https://dev.eclipse.org/mailman/listinfo/cross-project-issue s-dev

There is a lot of "developer noise" there, but also often some good "use=
r reports" of cross project issues, how to do the next simultaneous rele=
ase better, etc. While not clearly documented, it is supposed to be "a l=
ist for developers and users who encounter problems with Europa: can't d=
ownload, can't install, unable to make these two projects work together,=
etc.".

Of course, another complaint we hear occasionally is theres too many lis=
ts and newsgroups :(
Re: quality and stability of Europa [message #47509 is a reply to message #47389] Wed, 25 July 2007 14:43 Go to previous messageGo to next message
Eclipse Webmaster is currently offline Eclipse Webmaster
Messages: 439703
Registered: July 2009
Senior Member
Ed Merks wrote:
> For little humor, it's interesting to see how the User Assistance
> category is truncated to just after the second "s" of the second word in
> the tabular view. :-P

Hahaha

Hrm, perhaps we should modify the Bugzilla template to truncate at 9
characters ;)

D.

>
>
> David Williams wrote:
>> On Sun, 22 Jul 2007 19:29:23 -0400, Benson Margulies <bimjunk@basistech.com> wrote:
>>
>>
>>> Dear Eclipse Foundation,
>>>
>>>
>> [....]
>>
>>> Sincerely,
>>>
>>> Benson Margulies
>>> CTO Basis Technology Corp.
>>>
>>>
>>
>> Benson,
>>
>> Thanks for your comments. Some of the problems you report in your posts about WTP do sound serious, and am surprised I haven't heard of them before. I checked bugzilla and our webtools newsgroup and couldn't find any reports from you there, or about these bugs (with my quick scan) ... so, guess you know, it would be best to open bugs on issues you find. Any WTP specific issues should be addressed via bug reports or the webtools newsgroup, but seemed your post here to the Foundation newsgroup was to question the Foundation on the meaning, purpose, and process of the yearly coordinated releases. So, I'll make a few general comments on that.
>>
>> Perhaps the marketing messages mismatched reality (that is rare, though, right? :) but the primary emphasis of our yearly simultaneous releases is just that -- simultaneous release. And the theory is (or, should I say, the hope is) that by doing that, we will at least promote more "simultaneous use" which in years past was literally impossible, since various versions of various projects never did work together at all, except for some special cases at some very specific windows of time.
>>
>> Many of us were disappointed that more testing wasn't done with Europa as the milestones progressed, by the projects and by the community, and that many of us projects found it necessary to be making pretty big changes near the end, instead of a gradual smooth shutdown. So, we at Eclipse are hoping to improve that next time around. However, there actually were some pretty big "cross-project" bugs fixed, that would have not been possible to fix (to detect in time, that is) if we had not done the simultaneous release ... and, for good and bad, many of those worst bugs are never seen: good since they should not be seen, bad since it's hard for others to see the progress made ... meager as it might be.
>>
>> So, again just want to say thanks, we value all comments, positive and critical. The more specific comments you can make, in bugzillas, the better it will be, and the more you can participate throughout the release cycle, the better it will be. I know not everyone can afford to participate throughout a release cycle, and can just most easily make comments right after a release, but I think the success/failure of Ganymede (and Eclipse as a whole?) depends on that sort of on-going participation, so, I hope you, and other readers of this thread, can join in that sort of participation for Ganymede.
>>
>>
>
Re: quality and stability of Europa [message #47539 is a reply to message #47347] Sun, 29 July 2007 16:42 Go to previous messageGo to next message
Bjorn Freeman-Benson is currently offline Bjorn Freeman-Benson
Messages: 335
Registered: July 2009
Senior Member
Ed,
You've got it backwards. The release train *will* kick projects off the
train if they fail to follow the collective rules. And we will do a
release train without EMF if the EMF project is sufficient unreliable -
trust me on this. But the reason that we won't need to do this is the
same reason that projects are willing to depend on EMF: because EMF has
demonstrated that it is a reliable project. If EMF were not reliable and
helpful to the community, it would not be one of the core projects.

In other words, the reason EMF is trusted is because it has earned that
trust through merit. The reason that, for example, Eclipse Monkey is not
trusted is because it has not earned that trust. If Monkey had shown the
same community involvement and responsiveness, then it would be a +1
project too.

Your claim that EMF will not be kicked off the train because it is a +1
project is wrong: EMF is a +1 project because other people trust it not
to be kicked off the train.

- Bjorn

> And that's typically a hollow threat as I've
> pointed out to folks. You can kick off the caboose, but you can't kick
> off one of the locomotives; i.e., a release train without EMF isn't
> possible.
Re: quality and stability of Europa [message #47572 is a reply to message #47539] Sun, 29 July 2007 21:08 Go to previous message
Eclipse User
Originally posted by: merks.ca.ibm.com

Bjorn,

I'm sure we are in total agreement. I have no intention of ever failing
to follow the collective rules to see how that pans out because I
believe (and know for a fact) that it would be impossible to have a
release train without a reliable and fully functional version EMF. EMF
is not unique in this role as a component that other's rely on, but is a
good example of such a thing. I think we are so successful at Eclipse
because many of us---and I would dare say all of us---value our
community and want to be exemplary in what we deliver We have a great
bunch of people with the exactly this kind of attitude and that's what
makes Eclipse great!


Bjorn Freeman-Benson wrote:
> Ed,
> You've got it backwards. The release train *will* kick projects off
> the train if they fail to follow the collective rules. And we will do
> a release train without EMF if the EMF project is sufficient
> unreliable - trust me on this. But the reason that we won't need to do
> this is the same reason that projects are willing to depend on EMF:
> because EMF has demonstrated that it is a reliable project. If EMF
> were not reliable and helpful to the community, it would not be one of
> the core projects.
>
> In other words, the reason EMF is trusted is because it has earned
> that trust through merit. The reason that, for example, Eclipse Monkey
> is not trusted is because it has not earned that trust. If Monkey had
> shown the same community involvement and responsiveness, then it would
> be a +1 project too.
>
> Your claim that EMF will not be kicked off the train because it is a
> +1 project is wrong: EMF is a +1 project because other people trust it
> not to be kicked off the train.
>
> - Bjorn
>
> > And that's typically a hollow threat as I've
>> pointed out to folks. You can kick off the caboose, but you can't
>> kick off one of the locomotives; i.e., a release train without EMF
>> isn't possible.
Previous Topic:Eclipse Summit Europe Registration Opens
Next Topic:Platform Webinar
Goto Forum:
  


Current Time: Fri Aug 01 06:01:29 EDT 2014

Powered by FUDForum. Page generated in 0.03163 seconds