Ahmed Mohombe wrote:
I just can't
currently prioritize the time to investigate and solve them myself.
Than if none of the Velocity developers has time for the project, IMHO
it would
be fair to say and state that, so that everybody knows Velocity is "de
facto abandoned" (or "dormant" in Apache language if I'm not wrong).
There are actually 2 basic problems here. The first problem is the one
you mention, which is that development is basically stagnant. The other
related problem is that the Velocity committers are less than 100%
forthright when asked about the status of the project.
There is a complete electronic record of all the code that has been
committed in this project. If you were to scan this from 2002 to
present, you would probably see entire years in which basically no code
was committed. Nonetheless, last fall, Henning responded to one of these
questions about the status of the project, by saying that the project
was being actively developed and always had been. I don't like using the
'L' word, but it is my considered opinion that Henning was knowingly
telling a falsehood. At other times, Will Glass-Husain has made comments
about the status of the project that seem intentionally designed to
mislead the reader about the true state of the project.
Now, I would grant that some things have been done over the last few
years. However, I would also conjecture that a cold-blooded analysis of
all the forward movement in Velocity over the last 4 years would show
that it is an amount of work that a single motivated person could have
done in a matter of a few weeks, if that. Even taking into account that
the people involved are volunteers and so on, what has been achieved
over a time period of 4 years or more is incredibly thin.
Just consider how long it has taken for basic things to happen. Decimal
arithmetic in templates was introduced at some point. However, the
initial patch for this was contributed by Peter Romianowski in 2002. It
literally took years (!) of sterile debate to finally introduce this.
Somebody offered a patch for map literals around the same time. I
believe this is one of the new features in the as-yet unreleased
Velocity 1.5. Yes, a patch offered nearly 4 years ago, that was just
sitting there in the bug tracker.
Velocity's deficiencies in whitespace handling have been well known for
many years. A proposal was hashed out at some point in 2001. (This
proposal is approximately the way FreeMarker handles the issue.) To my
knowledge, there has been absolutely no work done in terms of
implementing said proposal.
The deficiencies in Velocimacros are well known and I won't detail them
all here because the message is already too long... Again, to the best
of my knowledge, there is no ongoing work to address Velocity's
deficient macro system.
I have looked at the Velolcity 2.0 RoadMap wiki page, and it is my
impression, that if all the features in that roadmap for a 2.0 version
were implemented, Velocity might be close to catching up with the state
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.
Asking from newbies to "fix themselfs" is not OK IMHO and not the right
signal IMHO.
I can imagine that this is one of the most important but "not spoken
because of netiquette" reason
so many moved to FreeMarker.
Well, I don't know of any rule of netiquette which should prevent people
from pointing out that an open source project is inactive. (OTOH, when
people claim that an inactive project is active and "always has been",
it may breach netiquette to point out that the people in question are
lying. Though, then again, telling falsehoods should also be considered
a breach of netiquette, IMO....)
Actually, the lack of ongoing development on Velocity has been a reason
frequently cited for people switching to FreeMarker. Other reasons vary
across the board. The better error reporting in FreeMarker was quite
important for both Webwork and Hibernate-tools. The Webwork people were
quite enamored of FreeMarker's support for JSP taglibs.
Other people mention the whitespace handling, the more powerful macro
system, built-in support for internationalization, serious XML
processing capabilities....
But it all boils down to a basic fact: Velocity development has
stagnated, and as a result, the product has fallen further and further
behind the state of the art.
If a user has to fix for himself every library is depending on than it
simply makes no sense to use
those libraries at all.
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. By overvaluing the
apache "brand", people often end up wasting a lot of time with inferior,
unmaintained tools -- when far better alternatives are available.
Jonathan Revusky
--
lead developer, FreeMarker project, http://freemarker.org/
Ahmed.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
