From: bob
Subject: Is UCW ready for production use?
Date: 
Message-ID: <1176755992.030480.188020@p77g2000hsh.googlegroups.com>
Hi,

I'm a Lisp newbie, I've done some hello-world web apps in a few
frameworks, and UCW looks really promising.  However, I heard that the
UCW code is changing pretty fast and it's not ready for production
use.  Does anyone find this to be an issue?  Is there a subset of
features that are already stable?  What other issues should I know
about?  Any alternative recommendations would be appreciated.

Thanks! :)

Bob

From: Vagif Verdi
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1176757845.763460.287770@y5g2000hsa.googlegroups.com>
Yes, it does change faster than i keep up.
And it has to be hosted on linux.
I ended up chosing huncehntoot.
Advantages:
1. Much simpler. You can pick it up easy and quick.
2. All process of creating a web site is documented in one place, and
it is not obsolete (like ucw tutorials)
3. Runs on windows (lispworks) and linux (lispworks, sbcl)
4. Does not require you to download and install all common-lisp.net
repository of projects ;)

Anf finally - ucw makes sence only if you are creating specificly html
+javascript web applications.
For rich web applications, all its power is useless, because paradigm
is completely different.

I'm using hunchentoot with Abobe Flex. ANd i do not use html
generation library at all. Only cl-json
From: bob
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1176920175.983947.190130@b58g2000hsg.googlegroups.com>
Thanks for all the advice!  I'll look deeper into hunchentoot.

In the mean time, it struck me how the various frameworks handle
session storage.  It've read the discussion between Marco Baringer and
Dave Roberts http://common-lisp.net/project/ucw/docs/html/rest/rest.html
, and it seems that UCW continuations are stored as a series of
closures, and for now we can only store them in the main memory.
Marco also mentioned that somehow serializing them is a bad idea.  If
it's really as bad as it seems, (sorry to go a little off topic) is it
possible to sit it on top of a distributed virtual memory over the LAN
on the OS level, so we can make use of the web servers' main memories
and local disks transparently?  I know this is done quite often in
supercomputing, and google seems to be doing something similar, but
I'm curious why it's rarely seen in beige box web farms (most cheap
setups seem to rely on app-level cache like memcached).

Thanks!

Bob
From: Drew Crampsie
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <462420ac$0$16321$88260bb3@free.teranews.com>
bob wrote:
> Hi,
> 
> I'm a Lisp newbie, I've done some hello-world web apps in a few
> frameworks, and UCW looks really promising.  However, I heard that the
> UCW code is changing pretty fast and it's not ready for production
> use.  

The ucw_dev code is very stable and i've been using it (and its
predecessors) in production for years.

The ucw_ajax branch is newer and has a lot of bells n whistles.. i
couldn't tell you much about stability though.

Does anyone find this to be an issue?  Is there a subset of
> features that are already stable?

The basic RERL is quite 'stable', but UCW is not really for lisp
newbies. Unless your application will really benefit from the 'modal'
nature of a continuation-based framework, i'd suggest going with
hunchentoot and one of the various template toolkits. It will be a lot
more familiar than trying to learn Common Lisp, CALL/CC, and UCW's
unique Request-Eval-Response loop all at the same time.

Having said that, when i started with lisp i jumped right into UCW and
never turned back. And it was a lot less 'stable' and a lot more 'not
ready for production' back then :).

> What other issues should I know
> about?  

UCW is going through a transitional phase in its development. We're in
the process of cleaning up the code, stabilizing the 'core' engine and
modularizing the rest. While this is going on, i don't think many of the
maintainers are going to pay much attention to _dev and _ajax, therefore
bitrot it likely to set in. That, and UCW requires a lot of libraries,
many from darcs/cvs etc and can be a major PITA to set up.

UCW's major features, continuations and components, are an incredible
way to model web applications, and i find it very productive, so don't
let any of this get you down.. if you really want to use UCW (and you
should it's really cool), the community around it is active enough that
you will be able to get the support you need.

Any alternative recommendations would be appreciated.

Hunchentoot is a good choice for a more 'traditional' approach. I've
also heard good things about webactions, though i'm not a fan of paserve
 myself.

hth,

drewc
> 
> Thanks! :)
> 
> Bob
> 

-- 
Posted via a free Usenet account from http://www.teranews.com
From: ·················@gmail.com
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1176813628.778708.299400@n76g2000hsh.googlegroups.com>
On Apr 17, 4:12 am, Drew Crampsie <·············@gmail.com> wrote:
> bob wrote:
> > Hi,
>
> > I'm a Lisp newbie, I've done some hello-world web apps in a few
> > frameworks, and UCW looks really promising.  However, I heard that the
> > UCW code is changing pretty fast and it's not ready for production
> > use.  
>
> The ucw_dev code is very stable and i've been using it (and its
> predecessors) in production for years.
>
> The ucw_ajax branch is newer and has a lot of bells n whistles.. i
> couldn't tell you much about stability though.
>
> Does anyone find this to be an issue?  Is there a subset of
>
> > features that are already stable?
>
> The basic RERL is quite 'stable', but UCW is not really for lisp
> newbies. Unless your application will really benefit from the 'modal'
> nature of a continuation-based framework, i'd suggest going with
> hunchentoot and one of the various template toolkits. It will be a lot
> more familiar than trying to learn Common Lisp, CALL/CC, and UCW's
> unique Request-Eval-Response loop all at the same time.
>
> Having said that, when i started with lisp i jumped right into UCW and
> never turned back. And it was a lot less 'stable' and a lot more 'not
> ready for production' back then :).
>
> > What other issues should I know
> > about?  
>
> UCW is going through a transitional phase in its development. We're in
> the process of cleaning up the code, stabilizing the 'core' engine and
> modularizing the rest. While this is going on, i don't think many of the
> maintainers are going to pay much attention to _dev and _ajax, therefore
> bitrot it likely to set in. That, and UCW requires a lot of libraries,
> many from darcs/cvs etc and can be a major PITA to set up.
>
> UCW's major features, continuations and components, are an incredible
> way to model web applications, and i find it very productive, so don't
> let any of this get you down.. if you really want to use UCW (and you
> should it's really cool), the community around it is active enough that
> you will be able to get the support you need.
>
> Any alternative recommendations would be appreciated.
>
> Hunchentoot is a good choice for a more 'traditional' approach. I've
> also heard good things about webactions, though i'm not a fan of paserve
>  myself.
>
> hth,
>
> drewc
>
>
>
> > Thanks! :)
>
> > Bob
>
> --
> Posted via a free Usenet account fromhttp://www.teranews.com

I just started a web project but i still didn't settled on web server/
framework,
(still evaluating UC, Hunchentoot and  KPAX )
 so could you point me what kind of sites us UCW good and
what kind of sites are better for Hunchentoot?
Some examples would be highly appreciated?

Pet

Pet
From: Drew Crampsie
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <4624dc3b$0$16381$88260bb3@free.teranews.com>
·················@gmail.com wrote:
>> Posted via a free Usenet account fromhttp://www.teranews.com
> 
> I just started a web project but i still didn't settled on web server/
> framework,
> (still evaluating UC, Hunchentoot and  KPAX )
>  so could you point me what kind of sites us UCW good and
> what kind of sites are better for Hunchentoot?

UCW for web applications, Hunchentoot for web sites, both for both. If
you install and play with them, you'll know what's good for what.

> Some examples would be highly appreciated?

Are you sure that you've got the punctuation right there!

drewc

> 
> Pet
> 
> Pet
> 

-- 
Posted via a free Usenet account from http://www.teranews.com
From: ·················@gmail.com
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1176891001.915740.298120@y5g2000hsa.googlegroups.com>
On Apr 17, 5:33 pm, Drew Crampsie <·············@gmail.com> wrote:
> ·················@gmail.com wrote:
> >> Posted via a free Usenet account fromhttp://www.teranews.com
>
> > I just started a web project but i still didn't settled on web server/
> > framework,
> > (still evaluating UC, Hunchentoot and  KPAX )
> >  so could you point me what kind of sites us UCW good and
> > what kind of sites are better for Hunchentoot?
>
> UCW for web applications, Hunchentoot for web sites, both for both. If
> you install and play with them, you'll know what's good for what.
>
> > Some examples would be highly appreciated?
>
> Are you sure that you've got the punctuation right there!

Let me ask my question little differently,
I've played with  huchentoot and kpax (ucw still waits it's turn)
and made a quasi-bank demo, i.e.  manipulating user acoounts rights,
"payments", calculating credit rating, etc through browser ,
which correct me if i'm wrong is kind of web application.
What are the advantages of making this kind of staff in ucw .

Also having a chance to look at a site made with ucw  would
make me understand what people are doing with ucw
more than a thousands of words. So could you please post
some links to web applications made with it.

thank you in advance

pet
From: tritchey
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1176990758.650075.18610@b58g2000hsg.googlegroups.com>
On Apr 18, 6:10 am, ·················@gmail.com wrote:
> So could you please post some links to web applications made with it.

We use ucw for <http://paragent.com>. You can click on the "DEMO"
button up at the top to get into a live sandbox of the application.
From: fireblade
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1177072615.263282.114020@o5g2000hsb.googlegroups.com>
On Apr 19, 3:52 pm, tritchey <········@gmail.com> wrote:
> On Apr 18, 6:10 am, ·················@gmail.com wrote:
>
> > So could you please post some links to web applications made with it.
>
> We use ucw for <http://paragent.com>. You can click on the "DEMO"
> button up at the top to get into a live sandbox of the application.

Impressive , how long are you developing  it?

bobi
From: tritchey
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1177076259.890352.264740@y80g2000hsf.googlegroups.com>
On Apr 20, 8:36 am, fireblade <·················@gmail.com> wrote:

> > We use ucw for <http://paragent.com>. You can click on the "DEMO"
> > button up at the top to get into a live sandbox of the application.
>
> Impressive , how long are you developing  it?
>
> bobi

We started working on it in anger around July of last year, and had
our first customers on it in November. There have been a series of
minor updates along the way, and we are getting ready to release a
major new version with integrated remote-desktop and a new ticket/help-
desk feature.

UCW has worked very well for us. There are two of us that work on it.
The other developer does most of the ucw stuff, so I am not the best
source of opinion on it. However, there is a pretty good group of
people that work on it, and someone always seems available on IRC on
the #lisp or #ucw channels to help out.
From: Holger Schauer
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <yxzveft94p6.fsf@gmx.de>
On 4976 September 1993, Drew Crampsie wrote:
> UCW is going through a transitional phase in its development. We're in
> the process of cleaning up the code, stabilizing the 'core' engine and
> modularizing the rest.

That's good to hear. Even better would be if this time somebody would
care about writing user documentation, too. 

Holger

-- 
---          http://hillview.bugwriter.net/            ---
"Ich habe neulich einen Emacs-Anhaenger gesehen, an einem
 Schluesselbund. Sehr characteristisch: Er war etwa hausgross und
 mehrere Tonnen schwer."
                  -- Kristian Koehntopp in de.comp.os.unix.discussion
From: Matthew D Swank
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <pan.2007.04.18.14.32.50.768993@c.net>
On Wed, 18 Apr 2007 15:45:25 +0200, Holger Schauer wrote:

> On 4976 September 1993, Drew Crampsie wrote:
>> UCW is going through a transitional phase in its development. We're in
>> the process of cleaning up the code, stabilizing the 'core' engine and
>> modularizing the rest.
> 
> That's good to hear. Even better would be if this time somebody would
> care about writing user documentation, too. 
> 
> Holger

Yes the doc could be better.  However, we're not talking Cells here.
Sometimes I think source-as-doc is the programming equivalent to an oral
tradition.  Whenever Kenny tells us to go read the PyCells doc, I have
this involved fantasy in which Python missionaries go visit the the Cells
camp and document their culture right before they're converted
and assimilated.  Kenny usually ends up being tried as a heretic.

I supposed this means I should get out more.

Matt 

-- 
"You do not really understand something unless you
 can explain it to your grandmother." — Albert Einstein.
From: Ties  Stuij
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1177054505.328009.50650@y5g2000hsa.googlegroups.com>
> Even better would be if this time somebody would
> care about writing user documentation, too.

http://trac.common-lisp.net/ucw/wiki/UcwIntro
Not entirely up to date i believe, but close enough. Written for
ucw_dev, but ucw_ajax isn't all that different. For both the current
ucw_dev and ucw_ajax practices, check the examples in the sources,
which on their own might just be all you need.

or

http://www.cliki.net/ucw%20Tutorial
From: Holger Schauer
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <yxzodlj8guh.fsf@gmx.de>
Hi Ties,

On 4979 September 1993, Ties Stuij wrote:
>> Even better would be if this time somebody would
>> care about writing user documentation, too.

> http://trac.common-lisp.net/ucw/wiki/UcwIntro

We had a discussion about that one before, I think, or at least on
some aspect of it.

> Not entirely up to date i believe, but close enough.

Actually, I'm just too lazy to write up a major overhaul
and send it to you. I think there is *a lot* of information
missing, mostly related to the flow of control in UCW.

Random stuff I had a quite a hard time to figure out:
- How do you set/read information from a subcomponent in a
  switching-container?
- How do you trigger an action directly from the lisp side?
- When exactly are render and refresh-component called when
  you're looking at some component that has subcomponents
  or called out to other components?
- The usefulness of :before and :after methods on render.
- When is it possible to answer/call at all?
- If you have some component that can be displayed on its own, but
  also be called from some other component, how do you determine 
  if and when to answer? (I resorted to checking whether the slot
  it.bese.ucw::calling-component is bound; note that the symbol isn't
  exported.)
- When/how (long) is the information in subcomponents that are
  (simple-)forms available?

And finally, the integration of flow of control and tal files is a
topic of it's own (that's where auxillary methods on render are really
helpful). Actually, I' still far from certain that if I were to write
that overhaul, it wouldn't be grounded on private conclusions and
assumptions -- after all, I know next to nothing about the inner
workings of UCW.

Please, don't take that as an attack on the tutorial, quite to the
contrary, I wouldn't have come anywhere without it. But it's not as if
there wasn't room for progress. And finally, I believe that UCW is
incomprehensible enough that it should provide documentation that goes
a large step further than a simple introduction.

Holger

-- 
---          http://hillview.bugwriter.net/            ---
"Ich habe neulich einen Emacs-Anhaenger gesehen, an einem
 Schluesselbund. Sehr characteristisch: Er war etwa hausgross und
 mehrere Tonnen schwer."
                  -- Kristian Koehntopp in de.comp.os.unix.discussion
From: Ties  Stuij
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <1177144682.857644.281150@l77g2000hsb.googlegroups.com>
> > Not entirely up to date i believe, but close enough.
>
> Actually, I'm just too lazy to write up a major overhaul
> and send it to you. I think there is *a lot* of information
> missing, mostly related to the flow of control in UCW.

Oh, no, i wasn't claiming i penned down anything conclusive, but
writing doc isn't my hobby i must say. The outset was to provide a
base. I figured one could figure out the rest by pressing down on m-.
a lot. And if you have a hard time figuring out one part, than that's
probably a nice part to add to the wiki.

/Ties
From: Friedrich Dominicus
Subject: Re: Is UCW ready for production use?
Date: 
Message-ID: <873b2uxkgz.fsf@flarge.here>
Holger Schauer <··············@gmx.de> writes:

>
> Random stuff I had a quite a hard time to figure out:
> - How do you set/read information from a subcomponent in a
>   switching-container?
> - How do you trigger an action directly from the lisp side?
> - When exactly are render and refresh-component called when
>   you're looking at some component that has subcomponents
>   or called out to other components?
> - The usefulness of :before and :after methods on render.
> - When is it possible to answer/call at all?
> - If you have some component that can be displayed on its own, but
>   also be called from some other component, how do you determine 
>   if and when to answer? (I resorted to checking whether the slot
>   it.bese.ucw::calling-component is bound; note that the symbol isn't
>   exported.)
> - When/how (long) is the information in subcomponents that are
>   (simple-)forms available?
I would ask that questions on the ucw  mailing list. IMHO there's way
to much guesswork needed to get anywhere, YMMV of course

Regards
Friedrich


--
Please remove just-for-news- to reply via e-mail.