Please, don't say "NOOOOO!" immediately. Read the mail carefully,
take a deep breath and think cold.
After working a couple of weeks in
the Metacello configurations and now with Damien's scripts to build Dev and Web
images, I came to the question:� Do we really need a PharoWeb
image?
I will give just my opinions of why having PharoWeb images is not
a good idea anymore
1) First of all, I don't like the PharoWeb name.
Because it is not Web in general, it is seaside. So, in my opinion, a better
name would be PharoSeaside. Maybe even new comers may think that the only
alternative is seaside. You know I have nothing against Seaside and that I even
use it several times. I would like to see AidaWeb images based in Pharo too.
2) Why to still have PharoWeb images if now Seaside images (at least the
one click) are based in seaside ?� In addition, when I develop in Seaside,
I use Seaside images, not PharoWeb, and I think most people do the
same.
3) Seaside images (or Lukas images for seaside) are MUCH more
stable than the PharoWeb images. Why? Several reasons:
�� 3.a) THEY
know the dependencies of the packages, the versions, how to set up, the order of
load, etc,� NOT us. THEY are the experts in seaside, not
us.
�� 3.b) THEY really test their images before make then public.
�� 3.c) We are "force" to release one release per month which has
no sense. How says that Seaside is released once a month? That's bad because it
means we are usually taking unstable versions. THEY know when a release should
be done and when a release is stable or not.
4) Now I created all the
Metacello configurations for the PharoDev image and I took me a lot of time.
Because I am not the developer of any of those projects. So, I need to
understand the dependencies, the order, the versions, etc. But why ??? If
Seaside people have already done that, don't duplicate efforts.
5) The
instability of PharoWeb images give to our public a bad impression. We waste all
of out efforts of Pharo in that image.
6) The idea of Pharo is to be the
vehicle. And that's exactly the Pharo role in the Seaside images. We don't need
to build PharoWeb images.
Even more, do we build PharoGlass images ? NOOOO.
Why not ? Because Dale already prepare them and as HE knows much more about
Glass than us, his images are better for Glass.
7) Tracking and managing
issues always take time. Even analyzing a bug, say "we won't fix" and close it.
With PharoWeb images we are having those kind of issues.
8) Our resources
are limited. Why not to use them in more important stuff ? For example, building
a better and more robust Dev image?
Here is my
proposal:
1) Completely remove the idea of PharoWeb image. Remove the
links from the website and do not create Web images any more.
2) In
http://www.pharo-project.org/pharo-download�
add a description explaining that now Seaside images are based in Pharo, so you
can directly download that image. We can even explain a bit about seaside, and
even if you want you can put a href called "PharoWeb" which links to the seaside
image. However, I would rather call it "PharoSeaside".
3) Tomorrow we can
have AidaWeb images based in pharo (please Janko!!)� so we can just do the
same and put the link to those images.
4) You may notice that Seaside images
are a bit "different" than our Dev images. This is mostly due to the
preferences, the theme and maybe some tools (ECompeltion instead of OCompletion,
etc..). So, what I suggest is to prepare a little script that can be
downloadable and those people who want a Seaside image but more "similar" to our
Dev images, just run that script.
So...what do you think? I would
really like to hear your opinion about
this.
Cheers
Mariano