Re: [studs-devel] rebuttal to blog post

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Yikes!  I was a little harsh on existing frameworks.  I didn't mean
ALL frameworks, I meant that there are alot.  I don't mean to
discredit quality work that has been done.  However, the fact still
remains, there is no standard base.  No standard way to import classes
or define a hierachy.

Dan

On Wed, 15 Sep 2004 11:37:19 -0400, Dan Allen <dan...@gm...> wrote:
> Well, let me answer some of those questions because I feel that I have
> good answers.  Oh, and don't worry about offending me.  I am a
> programmer and thus have extremely thick skin ;)
> 
> Why a Struts port in PHP?  Several reasons.  When I first got into MVC
> and read about Struts I thought to myself, that would be nice, if I
> had access to a servlet container, but I didn't (for deployment
> purposes, not for development).  I thought to myself, if I had this
> for PHP, I could make my website using it today.  Then, one day, when
> I do have access to a servlet container, all my ideas will translate.
> The transition would have been very smooth.  The PHP version is like
> the training wheels.  Now, I realize it is possible to get java
> hosting, but it would still require an ISP change.  Later on, I found
> even more reason since working with java servlets on a shared web
> hosts is a HUGE pain.  I already loved PHP, and I wanted to stick with
> it.
> 
> Okay, so I start working on the port, and it is hard.  There was a lot
> of work to go into it.  Why didn't I quit?  Several reasons again.  I
> had never taken a project from start to finish and made it a top-level
> sourceforge project.  This was my chance.  I needed it and it seemed
> interesting.  As I was working along, I found TONS of bugs in Tomcat
> and Struts, which I contributed back to the community.  I was reading
> all through the Tomcat code and it was amazing to see how it all
> worked.  The challenge was translating that to another language, which
> required a whole other level of understanding.  Today, I know the
> servlet container almost as well as the people who developed it.
> Ditto for Struts.  That understanding is without value.
> 
> Okay, that makes sense, but what benefit is it to PHP?  Look, PHP
> frameworks are shit.  Most of the code out there is shit.  Now, that
> is not to say that the programs around good in respect to what they
> are trying to accomplish.  phpBB code is horrid.  Why does this
> happen?  Well, even with all the frameworks in Java, at least they all
> share something in common.  They all run in a standard container
> environment, handling sessions the same, importing classes the same,
> do error handling the same.  Point is, there is a standard base.  PHP
> has no standard base.  PEAR is not it.  My hopes are to strike up
> something to this effect.  Studs is a prototype.  There are so many
> libraries that can be used as standalone, it probably makes up 75% of
> what you download.
> 
> Was it worth it?  You bet.  If only from the standpoint of a learning
> experience, it was invaluable.  I learned how to look at a project,
> swallow it, and make it happen.  The whole lifecycle.  I'm proud of my
> work.  I will use it.  Others can choose as they may, that is why it
> is open sourced.  Rails could EASILY be made using Studs MVC
> Framework+.  Just implement HttpServlet and go from there.  Maybe I
> will look into it.  Struts is not the end all solution to MVC, but it
> is a damn good starting point.
> 
> One last point.  You mentioned that it would be better to not have a
> configuration file and just use some sort of reflection on the action
> to get the class, etc.  I actually started that way with Studs and it
> is in the early code.  I went back and implemented struts-config.xml
> support because my goal was to create a stepping stone for PHP
> developers moving to Java.  I could easily make a framework which
> extends Studs and cuts out this configuration file dependency.
> 
> Oh, sorry, another last point.  Every PSP page is just a PHP page on
> drugs.  With no taglibs or JSTL syntax, it just compiles as itself.
> So in all reality, one could use Studs either was Model 1 or Model 2.
> You don't HAVE to create an Action, etc.
> 
> Dan
> 
> 
> 
> 
> On Wed, 15 Sep 2004 11:48:15 +0200, Balazs Fejes <fb...@gm...> wrote:
> > Hi,
> >
> > thanks for going through my points.
> > Can I ask you what you see as the benefits of this framework?
> > I think if you want to do it the "cleanly architected" way, you might
> > as well use Struts and JSTL.
> > I thought the benefit of PHP would be exactly the opposite -
> > simplicity (even at the expense of the architecture).
> > What are your thoughts on the more radical approach of Rails? You must
> > look at it when thinking about MVC frameworks. Maybe this would be a
> > worthwhile project to clone in PHP rather than Struts :-)
> >
> > I went through my post and it came off very agressive. Sorry about
> > that, but I felt like someone is putting in an incredible effort into
> > something that is not worth it.
> >
> > Best regards,
> > Balazs
> >
> >
> >
> >
> > On Tue, 14 Sep 2004 11:37:45 -0400, Dan Allen <dan...@gm...> wrote:
> > > In case you were interested, I am listening ;)
> > >
> > > ---------- Forwarded message ----------
> > > From: dan...@mo... <dan...@mo... >
> > > Date: Tue, 14 Sep 2004 08:33:16 -0700 (PDT)
> > > Subject: [studs-devel] rebuttal to blog post
> > > To: stu...@li...
> > >
> > > The following post showed up on the internet the other day about how Studs
> > > fails to be noteworthy.  While I disagree, one can never pass up the
> > > opportunity to improve on one's work.  So let's talk.
> > >
> > > "it seems that Studs takes the most annoying aspects of Struts
> > > development, and castrates the quick development process of PHP.
> > >
> > >     * web.xml required with Servlet registration, etc.
> > >     * PHP Server Pages??? custom View language
> > >     * tld files, property files
> > >     * struts-config.xml
> > >     * restarting/reloading is often neccessary??"
> > >
> > > 1. web.xml is a very important part of this project's philosophy because
> > > it is standard and it is familiar.  While I definitely like to think
> > > outside of the box, not all users will want to learn yet another
> > > configuration environment.  web.xml gives the flexibility of tools
> > > support, documentation and standarization.  Besides, most users won't even
> > > touch it, opting to take the standard configuration.  So I don't think it
> > > is an issue to the user.  It is more for the framework developer.
> > >
> > > 2. PHP Server Pages...yes.  No one can deny the simplicity of JSTL,
> > > especially in PHP, which lacks object dereferencing.  I tried initially to
> > > use PHP as the template language and it was FAR too verbose!  I believe
> > > Smarty is plain riduculous, because it is just another flavor of PHP.
> > > Phase is slick.
> > >
> > > 3. tld files are more or less superficial, mostly just for documentation
> > > purposes.  The developer doesn't even have to know about them, except when
> > > creating custom tags.  properties files are completely naturally in most
> > > programs anyway, so I am not sure what the complaint is here.  You need to
> > > configure the program SOMEHOW!
> > >
> > > 4. struts-config.xml...YES.  Standard, has tools support, very familiar to
> > > a large number of developers, lot's of online tutorials.  We are
> > > leveraging existing knowlege here folks.  However, I recognize Struts has
> > > its limitations and we are going to be going off on a tangent a bit to
> > > integrate them, perhaps tossing back the ideas to Struts.  DTDs are VERY
> > > limiting in this case because changing the DTD breaks the program when it
> > > doesn't have to.  I haven't broken the DTD yet, but it won't be long I
> > > bet.  We can talk.
> > >
> > > 5. restarting/reloading will go away.  PSP already reloads,
> > > struts-config.xml just needs to in a few releases.  I have plans for this,
> > > just haven't gotten to it.  I HATE restarting in J2EE, even though I
> > > understand the need for it in that environment.
> > >
> > > Studs aims to be simple while leveraging existing formats.  This is only
> > > inning one.  I wanted to create something that worked and was familiar.  I
> > > am open to extensions and improvements.  Hell, you can leverage all the
> > > tools save Studs to implement another popular MVC framework if you would
> > > like.  That is the benefit of the "+" part.  You can use the framework to
> > > do something totally different than Studs.  I just needed to port Struts
> > > for my own use, but feel free to jump in!
> > >
> > > One point I agree on is that PHP programs are ugly.  Many programs are
> > > just glorified perl scripts.  Studs seeks to add some maturity to the PHP
> > > community by laying out some standards so that programs can be shared more
> > > easily.  Studs definitely does not slow you down.
> > >
> > > Dan
> > >
> > > -------------------------------------------------------
> > > This SF.Net email is sponsored by: thawte's Crypto Challenge Vl
> > > Crack the code and win a Sony DCRHC40 MiniDV Digital Handycam
> > > Camcorder. More prizes in the weekly Lunch Hour Challenge.
> > > Sign up NOW http://ad.doubleclick.net/clk;10740251;10262165;m
> > > _______________________________________________
> > > studs-devel mailing list
> > > stu...@li...
> > > https://lists.sourceforge.net/lists/listinfo/studs-devel
> > >
> > >
> > > --
> > > Open Source Advocacy
> > > http://www.mojavelinux.com
> > >
> >
> >
> > --
> > Best regards,
> > Balazs
> >
> 
> 
> 
> 
> --
> Open Source Advocacy
> http://www.mojavelinux.com
> 

-- 
Open Source Advocacy
http://www.mojavelinux.com