Daniel Dekany wrote:
Sunday, May 7, 2006, 2:53:33 PM, Jonathan Revusky wrote:
[snip]
of FreeMarker of several years ago. That is how far behind the state of
the art this project is technically. In that vein, the 1.5 release date,
while important, is still too little, too late. Even a stable 1.5
release will not be remotely competitive with production/stable versions
of FreeMarker of three or even four years ago.
The question is, how do you measure if how competitive a project is.
I was only referring to the technical considerations. Velocity is not
currently competitive with FreeMarker technically. I mean they've got a
blue skies wish list for a hypothetical 2.0 version of Velocity that
does not contain anything that has not already been in FreeMarker for
several years.
If you measure it by how successful (nearly means, popular, widely
respected, famous) a project is, then it has little to do with the
technical qualities of the project. OK, a certain level of technical
quality must be reached, but after that... what matters is marketing,
i.e. the manipulation of what the mass thinks. Now, unlike Velocity,
FreeMarker is weak on the marketing field, and probably it always will
be.
Well, it isn't actually that Velocity has any "marketing". It simply
benefits from an "Apache effect". A very large number of people out
there place an amazingly high value on the "Apache brand".
I don't know exactly why this is so and I've wondered a lot about this.
People perceive the Apache brand as meaning something, and it inspires
some degree of confidence. OTOH, as far as I can see, there is no basis
whatsoever for having any particularly greater degree of confidence in
something from ASF than anything else throws at the wall to see if it
sticks. Just from observing what one sees here, for example, why would
anybody think that something being under the Apache umbrella is a
guarantee of anything at all? When asked about when much-needed
bug-fixes will be generally available in a release, the people
responsible for the project simply point out (correctly) that they are
unpaid volunteers and so on. Fair enough, but this is just the same
situation that you'd have with any fairly dormant project on
sourceforge. Why should anybody attribute any value to the "Apache brand"?
My thinking on this leads me to the idea that this has to do with a kind
of herding instinct. It's the idea that there is safety in numbers. And
that, actually, is rationally based to some extent (though not
absolutely, of course.) The thing is that the animals in the herd need
some point of reference to herd around, even a completely arbitrary
point of reference, and it seems that apache.org has evolved as such a
focal point.
Thus, to be competitive, let me mind you, that you (or... we)
should invest a *lot* into fixing all the design mistakes of
FreeMarker, and improve it's quality in general.
We should do that, but not specifically in order to compete with
Velocity. There is, currently, no technical competition with Velocity
going on.
Thus debating too
much on this list may not be a good way of using your time. You are
the main developer there after all.
Well, you, Daniel, originally became aware of FreeMarker because of
discussion on this list that I was involved in.
[snip]
Well, the thing is that these projects under a Jakarta/ASF umbrella get
a lot more attention and usage than they would otherwise, and surely,
this is precisely because people think that these projects are more
likely to be actively maintained and developed than other open source
projects. I think it's important that people should open their eyes and
realize that this is not particularly the case.
[snip]
Sure. For the sake of fullness, however, it should be mentioned that
FreeMarker also had and has problems with development inactivity.
That is definitely true. OTOH, our record, though far from perfect, is
far far better than the Velocity project. Bugs just sit around for years
here, even ones that look like they should be trivial to fix. For
example, one improvement (bug-fix really) in the latest version of
Velocity is that you can now write directives on multiple lines, i.e.
#set (
$x = $y
)
Clearly, VTL is a language that skips whitespace in directives and the
above should have worked. That bug was there for years, people were
aware of it, and Will, I think, finally fixed it a few months ago.
But this was just a glitch in the lexical specification in the grammar.
I mean, if something like this slipped in and was brought to our
attention, we would probably fix it on a same-day basis, and put out a
release maybe the next day with the fix. Again, if there is a bug of
this order in FreeMarker, it's less effort for me to just fix it than to
engage in any mailing list discussion....
It's
not better then Velocity if we ask how much is the maintenance
guaranteed.
Well, this is the thing. In neither case is there any mechanism that
guarantees the continuity, i.e. the future development and maintenance
of the project.
Of course, once you realize that you can only judge by track record, our
record has been much better on this.
Still, it has developed faster than Velocity... Why is
that? I think, that at the end of the day, it because of the mentality
of the main contributors. They really love to improve the stuff
(instead of seeking *excuses* if why a new feature goes against the
philosophy of the project, so they have to do nothing, or admit any
past mistakes). Just for the pure sake of joy. Some people there, are
genuinely interested in the template engine topic (yeah, there are
such perverts... :)), even if it they can seldom find time to lift
some heavy weight. It's that good-old hacker mentality (I don't mean
cracker here; nowadays hacker is often used as a synonym of that, who
knows why...), that was maybe there when the whole OS/FSF stuff have
started.
Well, I have been aware for some times that the ASF brand-name
projection has the perverse effect that people may want to get involved
with ASF projects for questionable reasons. I don't know exactly, but I
can't help but suspect that people think that getting involved with ASF
will further their careers somehow or other, provide something appealing
to put on their resumes....
I can only speculate about this, but the motive of many people is not
principally to hack code. That much seems clear.
Maybe I'm wrong, but I have the feeling that this mentality
is less and less significant at ASF (or maybe it never was
significant).
My sense of things is that in the early days, things were more loose and
informal and some things really could happen. Now, it's just so bogged
down in process and politics and so on that it looks impossible. I think
that projects get into ASF and just get bogged down in all this stuff.
Just look at what is happening with Webwork now that it is "merging"
with Struts. Amazing...
I feel it rather shows similarity to some kind of
business concern. (Jonathan had a good blog entry about this:
http://freemarker.blogspot.com/2006/02/some-comments-on-joining-jakarta-or.html)
If I'm right with my feelings, then I think that for most of the young
ASF fan people this thing is bad.
Well, I am convinced that the net effect of all of this on the "open
source ecology", as it were, it pretty negative.
They are in most cases not the part
of the group who has benefits from this. <sarcasm>Unless, of course,
they win a plasma TV by trying Geronimo.</sarcasm>
<LOL>
Well, I think that Geronimo is a flop basically, regardless of how many
plasma TV's they gave away.
Regards,
Jonathan Revusky
--
lead developer, FreeMarker project, http://freemarker.org/
FreeMarker group blog, http://freemarker.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
