Adrian Tarau wrote:
Hello Jonathan,
/"Well, in general, once you want to do anything moderately complex with
velocimacros, the thing breaks because it's.... *junk.* :-)"/
It's not nice to say about a different library(a "competitor") "it's
junk", even if the library is not as good as yours(I'm not sure if you a
commiter or just a user of FreeMarker).
What's not nice and a clear violation of nettiquette, as well as basic
common sense, is to respond to a private message (one that was clearly
meant to be private) in a public forum.
Well... okay, I guess you just don't know that. If you don't know, you
don't know. However, it stands to reason that if you don't know that,
there are may be plenty of other basic things you don't know.
So, first of all, there is no onus on me, particularly in a private
note, to pretend that I think that Velocity is anything other than
obsolete junk.
I agree, velocity has some weak parts, but this doesn't mean is "junk".
Well, that's debatable. There's no clear definition of what "junk"
means. A key characteristic of junk, at least most junk, is that, at one
point in the past, it was something of value. Some 386 or 486-based PC
lying around in someone's basement is junk, but at some point in the
past, it was a highly valued state of the art piece of equipment.
Getting back to niceness, since you accuse me of not being nice, one
thing that wouldn't be nice would be to sell that old 386 or 486 to
someone who had no knowledge of computers and represent that it was
something that was state of the art and so on.
I'm not pro Velocity and against FreeMarker of vice versa, and I am glad
for the existence of projects like Velocity or FreeMarker(all the Apache
projects, etc).
Besides nettiquette, a whole aspect of this you don't seem to understand
is that projects are really only useful if they are approximately
competitive with the state of the art in their space. I could write a
simple text editor or database this weekend and start announcing it on
all the appropriate forums. However, it would be a complete and utter
waste of time. Not only would I be wasting my own time (no big deal) but
I would trying to get people to waste their time as well. When things
like emacs and jedit exist, nothing I could write in short order would
be remotely of any interest to anybody. Look, it takes a certain ego and
nerve to announce your work to fellow professionals on places like
freshmeat.net and theserverside.com etcetera. To do this takes a certain
ego. You have to believe that, what with all the things out there that,
people really should take the time to look at what you've done. It has
to be broadly competitive with the state of the art.
So let's play nice...You can post articles about "FreeMarker is better
that Velocity" but do it with professionalism.
The whole idea that I, in the name of "professionalism" have to pretend
that the emperor is wearing any clothes, particularly in a private note,
is just absurd. Even in public there is no particular need for me to do
so. You may have noticed that people who review movies online and
elsewhere do not feel obliged to tell you that every movie that comes
out is actually good. And likewise, I am under no obligation to say that
every software tool out there is good. I don't consider Apache Velocity
to be something of high quality, to say the least. Not only is it
lacking in features that really should be considered by now to be basic
to templating in the web space, it embodies numerous first order
mistakes in design and implementation that, in over 7 years of
existence, nobody has put in the effort to remedy.
In my considered opinion, it is naive and misguided to feel gratitude to
ASF for something like Velocity. I believe that if Apache were run in a
way that really reflects its charter, to be of benefit to the overall
developer community, something in the state of Velocity would be labeled
as abandonware. Basically, the maintainers would have been put on notice
that they had to either get the thing in a state that it is
approximately competitive in its space, or the front page of the project
would say something like "Hi, I'm an abandoned piece of software. If you
want to adopt me, here is the procedure to do so." If Velocity and the
other many projects in the same state were labelled that way, it would
be of much greater benefit to the larger community. First of all, people
would be dissuaded from investing energy into something that is obsolete
and unmaintained. Also, people who actually do want to do the work of
maintaining the project would have a chance to do so.
But alas, obviously this is not the case. And specifically, that is why
so many software projects out there have initially used Velocity (based
on the belief that it was approximately competitive and being
maintained) and later had to switch to another tool, typically
FreeMarker. Some of the best known open source projects in the java
world went this route -- Hibernate, Netbeans, WebWork/Struts 2.
Obviously very many valuable developer man-hours were squandered this
way because Velocity was advertised as an active cutting edge project
when it wasn't.
I document this in a blog entry. See:
http://freemarker.blogspot.com/2007/12/velocity-of-freemarker-looking-at-5.html
PS . My framework(yet another java framework??? :) ) allows me to choose
between any template engine without modify any Java code so it is
transparent for me in general. This situation with calling dynamically a
macro is the first exception in years, usually I have anything I need in
Velocity, I don't need to think about another library(not that the
Velocity is the only good template engine, but I got used with it).
What you are saying is a big nonsequitir really. If I only used the 386
or 486 computer I mentioned above (along with applications of that
vintage) I would probably think it was great, assuming that I had no
exposure to the current state of the art in personal computing. Sure,
it's probably true that you have not felt that features were missing
from Velocity. Similarly people who have only used procedural
programming languages like C or FORTRAN or whatever, may feel that they
are not missing any of the features in more modern object-oriented
languages. However, once they do make the migration, they realize that
there are many far more elegant maintainable solutions to problems that
are now easy to implement that were not really possible before. In any
case, we see from many of these threads that other people are feeling
the lack of basic templating features -- macros with an associated
block, optional/default parameters, and so on.
In any case, I imagine you (and some other people) won't be happy about
this note I have written. However, you do seem to be suffering from some
very basic misconceptions about.... well... a whole bunch of things. I
write this on the forum, not just because you replied to me here, but
also because surely other people are suffering from similar
misconceptions, and if you don't take what I'm saying to heart, maybe
somebody will, so I won't be just venting. It will have served some purpose.
Jonathan Revusky
--
lead developer, FreeMarker project http://freemarker.org
Velocity or FreeMarker: Looking at 5 Years of Practical Experience
http://freemarker.blogspot.com/2007/12/velocity-of-freemarker-looking-at-5.html
Jonathan Revusky wrote:
Adrian Tarau wrote:
I've always used #if to implement the 'switch' but I think, even for
3-4 conditions, the template will look cleaner.
Instead of
#if('renderLabel' == $macroToCall)
#renderLabel($component)
#elseif('renderInput' == $macroToCall)
#renderInput($component)
#elseif(...)
...
#end
we will have
#call($macroToCall $component).
This kind of thing is trivial in FreeMarker. For example, suppose you
had:
<#assign macroHash = {'renderLabel' : labelMacro, 'renderInput' :
renderMacro, .... >
and then, supposing you have an action string, like suppose:
[#assign action = 'renderLabel']
then you could invoke the macro via:
<@macroHash[action] component />
The thing is that macros in FreeMarker are variables, and can be in
hashes or assigned to variables or whatever, and also the foo in
<@foo/> to invoke the macro can be any arbitrary expression.
So, for example, suppose the macro you want to invoke is in the string
macroName, you could invoke it via:
<@.vars[macroName] component/>
(.vars is a special built-in hash that contains the variables
available in the template and since macros are variables as well,
.vars[macroName] is the macro with the name macroName and it can be
invoked this way, or you could create a variable.
<#assign myMacro = .vars[macroName]>
and invoke it via:
<@myMacro component/>
Right below this, Mr. Van Bergen mentions Anakia, which is an add-on
to Velocity for processing XML. He neglects to mention that FreeMarker
provides similar XML processing functionality (though the
implementation is much more complete, since it supports XML
namespaces, for example) as part of its core feature set. Declarative
XML processing is supported in FreeMarker via the #visit and #recurse
directives, which are core directives in the FreeMarker language. One
would infer from what the article says that XML processing is a point
in favor of Velocity, when, really, quite the opposite is the case.
The XML processing functionality available for Velocity is add-ons
like Anakia and DVSL that are basically abandonware, where the XML
processing support in FreeMarker is a core part of the product, and is
clearly supported.
Well, in general, once you want to do anything moderately complex with
velocimacros, the thing breaks because it's.... junk. :-)
Here is a blog entry I wrote regarding some of this sort of thing:
http://freemarker.blogspot.com/2007/12/velocity-of-freemarker-looking-at-5.html
Jonathan Revusky
--
lead developer, FreeMarker project, http://freemarker.org/
In case of a value outside the 'domain' you will get an exception
"Macro not found ....".
We could have even a directive which will simulate the switch default
branch.
#callWithDefault($macroToCall $defaultMacro ....) - of course the
name should be shorter.
Christopher Schultz wrote:
Adrian,
Adrian Tarau wrote:
I have the following problem : I would like to call a macro but the
macro
name must be a variable.
This is more of a question for the user's list, not the dev list. In
the future, please post there.
Ex: instead of #renderLabel($component) to have #call("renderLabel"
$component) - of course "renderLabel" can be any (existing) macro
name.
How many possibilities can you have for $component? Are they
unlimited, or constrained to maybe 5 possibilities? I'm wondering
because you could easily do it like a switch:
#if('renderLabel' == $macroToCall)
#renderLabel($component)
#elseif('renderInput' == $macroToCall)
#renderInput($component)
#elseif(...)
...
#end
-chris
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
