[Pharo-project] ZnServer - Swazoo - Comanche
Hi Sven, People: I want to ask about ZnServer, is focused in have at some point all the features of Swazoo or Comanche and being itself a complete web server? Or the goal is only to be a default implementation, as the comment says? Thanks. Germán.
Hello Germán, On 19 Jun 2011, at 16:14, Germán Arduino wrote:
Hi Sven, People:
I want to ask about ZnServer, is focused in have at some point all the features of Swazoo or Comanche and being itself a complete web server?
Or the goal is only to be a default implementation, as the comment says?
Thanks. Germán.
From my point of view, the goal of Zn is to be a framework implementing the HTTP protocol and a library on top of that offering client and server functionality based mainly on RFC 2616.
The server part is kept relatively simple: the server does multithreading, connection handling, request parsing and response writing and then asks its delegate to do the work, using #handleRequest: which is given a ZnRequest as argument and is expected to return a ZnResponse. The code doing the #handleRequest can be simple, like in the examples that are included, but could become arbitrarily complex, like Apache. Anyone is free to build on top of Zn whatever they think makes sense, most of this will probably be more application specific and hence out of the scope of Zn. The Seaside adaptor is a good example: it is an add-on. Seaside expects very little from its servers and Zn is offering slightly more than needed. Still, the Seaside experience is quite rich to say the least. I don't consider Swazoo or Comanche competition, they offer more features but also specific API's. Another goal of Zn is understandability: I would like the Zn code to be readable by anyone interested in HTTP. Sven
Thanks Sven, very clear explanation. Germán. 2011/6/19 Sven Van Caekenberghe <sven@beta9.be>:
Hello Germán,
On 19 Jun 2011, at 16:14, Germán Arduino wrote:
Hi Sven, People:
I want to ask about ZnServer, is focused in have at some point all the features of Swazoo or Comanche and being itself a complete web server?
Or the goal is only to be a default implementation, as the comment says?
Thanks. Germán.
From my point of view, the goal of Zn is to be a framework implementing the HTTP protocol and a library on top of that offering client and server functionality based mainly on RFC 2616.
The server part is kept relatively simple: the server does multithreading, connection handling, request parsing and response writing and then asks its delegate to do the work, using #handleRequest: which is given a ZnRequest as argument and is expected to return a ZnResponse.
The code doing the #handleRequest can be simple, like in the examples that are included, but could become arbitrarily complex, like Apache.
Anyone is free to build on top of Zn whatever they think makes sense, most of this will probably be more application specific and hence out of the scope of Zn.
The Seaside adaptor is a good example: it is an add-on. Seaside expects very little from its servers and Zn is offering slightly more than needed. Still, the Seaside experience is quite rich to say the least.
I don't consider Swazoo or Comanche competition, they offer more features but also specific API's.
Another goal of Zn is understandability: I would like the Zn code to be readable by anyone interested in HTTP.
Sven
Hi guys, For Swazoo it would be fine to merge its HTTP framework with something else and here I see two possibilities, one from WebClient and another from Zinc. That way we can achieve the same HTTP library for both Swazoo and Zinc or WebClient tools. And later for web frameworks too, no need to transform requests and responses from/to framework specific ones as it is a case now. Main goal of Swazoo is to serve one or more virtual websites directly to the public Internet as optimal and reliable as possible, next goal is wide cross-dialect portability. Swazoo is doing that job reliably for 11 years now, so it will certainly stay there to continue. Just that it would be nice to improve it from time to time and having now Zinc and WebClient on hand it is a just nice opportunity for thinking about such a merge. Best regards Janko S, Sven Van Caekenberghe piše:
Hello Germán,
On 19 Jun 2011, at 16:14, Germán Arduino wrote:
Hi Sven, People:
I want to ask about ZnServer, is focused in have at some point all the features of Swazoo or Comanche and being itself a complete web server?
Or the goal is only to be a default implementation, as the comment says?
Thanks. Germán.
From my point of view, the goal of Zn is to be a framework implementing the HTTP protocol and a library on top of that offering client and server functionality based mainly on RFC 2616.
The server part is kept relatively simple: the server does multithreading, connection handling, request parsing and response writing and then asks its delegate to do the work, using #handleRequest: which is given a ZnRequest as argument and is expected to return a ZnResponse.
The code doing the #handleRequest can be simple, like in the examples that are included, but could become arbitrarily complex, like Apache.
Anyone is free to build on top of Zn whatever they think makes sense, most of this will probably be more application specific and hence out of the scope of Zn.
The Seaside adaptor is a good example: it is an add-on. Seaside expects very little from its servers and Zn is offering slightly more than needed. Still, the Seaside experience is quite rich to say the least.
I don't consider Swazoo or Comanche competition, they offer more features but also specific API's.
Another goal of Zn is understandability: I would like the Zn code to be readable by anyone interested in HTTP.
Sven
-- Janko Mivšek Svetovalec za informatiko Eranova d.o.o. Ljubljana, Slovenija www.eranova.si tel: 01 514 22 55 faks: 01 514 22 56 gsm: 031 674 565
participants (3)
-
Germán Arduino -
Janko Mivšek -
Sven Van Caekenberghe