I want to write a web framework. Which server should I use?

I wrote a really long email detailing why I think forking Seaside is a fool's errand, but I'm going to eschew that. If folks want to fork it, okay. If folks want to try to build a compat layer (huge amount of work, both in terms of code and in terms of liaising to the Seaside community) I'd give you my "doing it right" thumbs up, and bow to your bravery, also muttering something vague in clipped tlhIngan. That's all I'm gonna say though. I don't need to write a huge email to justify why I'm saying those two things, I don't think.

While you're all doing *that* though, I'd like to start actually serving some web. I have an idea about how I'd like to do it, too, and it meshes well with the design goals behind Cuis.

I've always envisioned Cuis being the best Smalltalk thing to run on the web frontend, because it's simple, clean, and fast. 

I know there's been some effort around here to get various web servers working. Can anyone here recommend which is the most reliable/usable, currently? Because I've had this web framework in my head for a year or three, and maybe now's the right time to implement it (if I'd done it when I thought of it, and didn't port it to other dialects, Juan would have been like "that's pretty cool." And then the conversation would likely have ended. Now I've got people wanting to talk web frameworks!)

TIA,

Casey

_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org

Hi Casey!

Your mail arrives in the same day that I was wondering aproximately the same things and, also, just a few time after I made some decisions.

I agree with the fact that the port of Seaside have the drawbacks that you exactly described. But I not think too much on it because I'm not really interested in Seaside for a couple of reasons (It's the antipodes of the simplicity that we want in Cuis, for example).

But I did other several mistakes, trying to port web tools to Cuis, for example invest a lot of time in too much things instead of select only the best options to focus there my efforts.

But some things only become revealed after some time messing with them, for example: I wanted to port Iliad, because is light and includes Ajax by default, but after working a bit on it seems that also have dependences of the style of Grease and really I've not received too much responses in the Iliad list. And, of course, I do not want to end as you says, with a fat compatibility layer, to be another Squeak or Pharo.

By other hand I received a lot of support from Janko and we have Aida working in Cuis and just today I started talking with Janko to port the latest version (Seems that with the experience of Janko in several Smalltalk this job should take only several hours). 

And, of course, I do not want to end as you says, with a fat compatibility layer, to be another Squeak or Pharo.

To get Aida working I ported Sport and Swazoo and I'm now the Sport/Swazoo maintainer for Cuis, then I think that this is the web server to use, because, also today, I was working a bit in Komanche, but....it's (again) a lot of work to use when? by who? The same I could apply to other ports that I've in progress (port is not only the initial port, also means MAINTAIN the port, adapt new versions, fix bugs, etc).

Then, I think today the same than you comment, "why not develop a framework in Cuis if needed, instead of deal with a lot of ports". But I do not have the time now....but I can help maintaining Sport/Swazoo if your decides to work in this project. 

And in this line of thinking I made a decision:

I will maintain:

- Sport

- Swazoo

- Aida (with the help of Janko)

I will abandon (by now) my work on:

- Iliad

- Komanche

- HttpView2

- Zinc (except that I can receive some sort of sponsorization)

Well, I hope I was clear and that my comments serve to your work

2013/10/1 Casey Ransberger <[hidden email]>

I wrote a really long email detailing why I think forking Seaside is a fool's errand, but I'm going to eschew that. If folks want to fork it, okay. If folks want to try to build a compat layer (huge amount of work, both in terms of code and in terms of liaising to the Seaside community) I'd give you my "doing it right" thumbs up, and bow to your bravery, also muttering something vague in clipped tlhIngan. That's all I'm gonna say though. I don't need to write a huge email to justify why I'm saying those two things, I don't think.

While you're all doing *that* though, I'd like to start actually serving some web. I have an idea about how I'd like to do it, too, and it meshes well with the design goals behind Cuis.

I've always envisioned Cuis being the best Smalltalk thing to run on the web frontend, because it's simple, clean, and fast. 

I know there's been some effort around here to get various web servers working. Can anyone here recommend which is the most reliable/usable, currently? Because I've had this web framework in my head for a year or three, and maybe now's the right time to implement it (if I'd done it when I thought of it, and didn't port it to other dialects, Juan would have been like "that's pretty cool." And then the conversation would likely have ended. Now I've got people wanting to talk web frameworks!)

TIA,

Casey

_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org

--

Saludos / Regards,
Germán Arduino
www.arduinosoftware.com

_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org