Catching up with Rails

You may or may not know that I have a local branch in my SVK
repository in which I'm trying to make typo work with edge
rails. 

It's, ah, fun.

I've reached the point where I am seriously considering
removing a
huge amount of code, leaving only the basic article
permalinks and
index/search pages as they are, then slowly adding (probably
starting
with posting a new article via the web) back in ways that
play well
with the new Rails Way; otherwise we end up pissing into the
wind.

This may take a while. But I'm not entirely sure it's a bad
idea. We're carrying a great deal of cruft in the typo
codebase where
it turns out that we've zigged before rails zagged.

My initial roadmap for this approach is looking like:

* Strip out everything but the basic blog display.
* Think about URL design for everything but permalinks,
ideally moving
  to a RESTful structure. The current /articles tree doesn't
really
  play well with the resourceful approach, but that's fine -
it's a
  useful archive structure when you're looking at the blog
as a blog.

  Currently I'm thinking in terms of resources like:

  /entries  => Articles treated as Resources
  /comments 
  /trackbacks
  /feedback => conflating comments and trackbacks, useful
for spam
               stopping
  /sidebars
  /pages
  /sessions => RESTful authentication
  /categories
  /tags
  /categorizations
  /taggings
  /users
  /rules    => Spam policy
  /blogs


  But I expect this to be a process of discovery. 

* Add comments + Akismet based spam checking
* Add RESTful authentication.
* Add article/page posting 
...

My thinking is that most features should go back in quite
quickly and
we'll end up with a rather more robust application that gets
to hand
more stuff off to Rails than we're currently doing. We're
more and
more out of step with the framework and that's causing us
pain.

-- 
Piers Cawley <pdcawleybofh.org.uk>
http://www.bofh.org.uk/
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

I say do it.  Every good project needs a good ground-up
rewrite now and 
then, and I think typo is past due.  There's a lot of cruft
in typo that 
just needs to go.  Can anyone say, AJAX Sidebar activation?

Matt



On Thu, 28 Sep 2006, Piers Cawley wrote:

> You may or may not know that I have a local branch in
my SVK
> repository in which I'm trying to make typo work with
edge rails.
>
> It's, ah, fun.
>
> I've reached the point where I am seriously considering
removing a
> huge amount of code, leaving only the basic article
permalinks and
> index/search pages as they are, then slowly adding
(probably starting
> with posting a new article via the web) back in ways
that play well
> with the new Rails Way; otherwise we end up pissing
into the wind.
>
> This may take a while. But I'm not entirely sure it's a
bad
> idea. We're carrying a great deal of cruft in the typo
codebase where
> it turns out that we've zigged before rails zagged.
>
> My initial roadmap for this approach is looking like:
>
> * Strip out everything but the basic blog display.
> * Think about URL design for everything but permalinks,
ideally moving
>  to a RESTful structure. The current /articles tree
doesn't really
>  play well with the resourceful approach, but that's
fine - it's a
>  useful archive structure when you're looking at the
blog as a blog.
>
>  Currently I'm thinking in terms of resources like:
>
>  /entries  => Articles treated as Resources
>  /comments
>  /trackbacks
>  /feedback => conflating comments and trackbacks,
useful for spam
>               stopping
>  /sidebars
>  /pages
>  /sessions => RESTful authentication
>  /categories
>  /tags
>  /categorizations
>  /taggings
>  /users
>  /rules    => Spam policy
>  /blogs
>
>
>  But I expect this to be a process of discovery.
>
> * Add comments + Akismet based spam checking
> * Add RESTful authentication.
> * Add article/page posting
> ...
>
> My thinking is that most features should go back in
quite quickly and
> we'll end up with a rather more robust application that
gets to hand
> more stuff off to Rails than we're currently doing.
We're more and
> more out of step with the framework and that's causing
us pain.
>
>

-- 
------------------------------------------------------------
--------------
Matt Rose     mattrosefolkwolf.net    Visit my blog!  http://folkwolf.net
He's an immortal dishevelled boxer with acid for blood.
She's a
bloodthirsty hip-hop museum curator with a knack for
trouble. They
fight crime!
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

Matt Rose <mattrosefolkwolf.net> writes:

> I say do it.  Every good project needs a good ground-up
rewrite now and 
> then, and I think typo is past due.  There's a lot of
cruft in typo that 
> just needs to go.  Can anyone say, AJAX Sidebar
activation?

Actually, that's fixable. But it's been problematic because
it really
didn't sit well with the workings of the sidebar object;
we're
roundtripping to the database far too often when we
could/should be
letting the browser retain the information. I've been
planning a
rewrite for about a yonk though.

The other option (which makes a great deal of sense for some
themes)
is to just have sidebar partials which you can edit in the
admin
interface (probably using liquid rather than rhtml). The
idea being
that it'd make it a *great* deal easier to have multiple
sidebars.

I'll see what I find out as I do the work though.

-- 
Piers Cawley <pdcawleybofh.org.uk>
http://www.bofh.org.uk/
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

I'm kind of conflicted about this.  On one hand, the end
result will
probably be great.  On the other hand, this is going to mean
a long
period of instability and random user complaints while we
finish
migrating everything.  And it'll also remove a bunch of code
that I
more or less understand and replace it with new code that
I'll have to
learn.

Perhaps we should consider finishing up 4.1 first and then
start on
this?  When is Rails 1.2 due?


Scott

On 9/28/06, Piers Cawley <pdcawleybofh.org.uk> wrote:
> You may or may not know that I have a local branch in
my SVK
> repository in which I'm trying to make typo work with
edge rails.
>
> It's, ah, fun.
>
> I've reached the point where I am seriously considering
removing a
> huge amount of code, leaving only the basic article
permalinks and
> index/search pages as they are, then slowly adding
(probably starting
> with posting a new article via the web) back in ways
that play well
> with the new Rails Way; otherwise we end up pissing
into the wind.
>
> This may take a while. But I'm not entirely sure it's a
bad
> idea. We're carrying a great deal of cruft in the typo
codebase where
> it turns out that we've zigged before rails zagged.
>
> My initial roadmap for this approach is looking like:
>
> * Strip out everything but the basic blog display.
> * Think about URL design for everything but permalinks,
ideally moving
>   to a RESTful structure. The current /articles tree
doesn't really
>   play well with the resourceful approach, but that's
fine - it's a
>   useful archive structure when you're looking at the
blog as a blog.
>
>   Currently I'm thinking in terms of resources like:
>
>   /entries  => Articles treated as Resources
>   /comments
>   /trackbacks
>   /feedback => conflating comments and trackbacks,
useful for spam
>                stopping
>   /sidebars
>   /pages
>   /sessions => RESTful authentication
>   /categories
>   /tags
>   /categorizations
>   /taggings
>   /users
>   /rules    => Spam policy
>   /blogs
>
>
>   But I expect this to be a process of discovery.
>
> * Add comments + Akismet based spam checking
> * Add RESTful authentication.
> * Add article/page posting
> ...
>
> My thinking is that most features should go back in
quite quickly and
> we'll end up with a rather more robust application that
gets to hand
> more stuff off to Rails than we're currently doing.
We're more and
> more out of step with the framework and that's causing
us pain.
>
> --
> Piers Cawley <pdcawleybofh.org.uk>
> http://www.bofh.org.uk/
> _______________________________________________
> Typo-list mailing list
> Typo-listrubyforge.org
> http:
//rubyforge.org/mailman/listinfo/typo-list
>
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

"Scott Laird" <scottsigkill.org> writes:

> I'm kind of conflicted about this.  On one hand, the
end result will
> probably be great.  On the other hand, this is going to
mean a long
> period of instability and random user complaints while
we finish
> migrating everything.  And it'll also remove a bunch of
code that I
> more or less understand and replace it with new code
that I'll have to
> learn.

Heh. I've just thrown up an experimental/restful branch with
my work
so far. Tests are passing with the exception of one that
seems to be a
rails bug, and the only deprecation warning we're getting
looks like
being another case of Rails making an unwarranted assumption
-- we
don't ever call the function it's complaining about.

> Perhaps we should consider finishing up 4.1 first and
then start on
> this?  When is Rails 1.2 due?

What are we lacking for 4.1? 

And I'm not sure when 1.2 is due, but the noises at the
European Rails 
Conf seemed to to imply it was coming real soon now.

-- 
Piers Cawley <pdcawleybofh.org.uk>
http://www.bofh.org.uk/
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

On 28-Sep-06, at 12:17 PM, Piers Cawley wrote:

> Matt Rose <mattrosefolkwolf.net> writes:
>
>> I say do it.  Every good project needs a good
ground-up rewrite  
>> now and
>> then, and I think typo is past due.  There's a lot
of cruft in  
>> typo that
>> just needs to go.  Can anyone say, AJAX Sidebar
activation?
>
> Actually, that's fixable. But it's been problematic
because it really
> didn't sit well with the workings of the sidebar
object; we're
> roundtripping to the database far too often when we
could/should be
> letting the browser retain the information. I've been
planning a
> rewrite for about a yonk though.
>

Is this why my new sidebar goes to the database to grab the 

information twice?

I've been puzzling this one over for quite some time.  I
have a  
sidebar that's pretty much complete now, however, it takes
well over  
30 seconds to load.   When I look at the logs, it's because
the  
sidebar goes back to the database and loads the feeds and
builds the  
feed, which is a pretty time consuming thing to do, and does
it twice.

Why is this?

Matt


> I'll see what I find out as I do the work though.
>
> -- 
> Piers Cawley <pdcawleybofh.org.uk>
> http://www.bofh.org.uk/
> _______________________________________________
> Typo-list mailing list
> Typo-listrubyforge.org
> http:
//rubyforge.org/mailman/listinfo/typo-list

_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

What's the URL for it?

Matt


On 28-Sep-06, at 7:24 PM, Piers Cawley wrote:

> "Scott Laird" <scottsigkill.org> writes:
>
>> I'm kind of conflicted about this.  On one hand,
the end result will
>> probably be great.  On the other hand, this is
going to mean a long
>> period of instability and random user complaints
while we finish
>> migrating everything.  And it'll also remove a
bunch of code that I
>> more or less understand and replace it with new
code that I'll  
>> have to
>> learn.
>
> Heh. I've just thrown up an experimental/restful branch
with my work
> so far. Tests are passing with the exception of one
that seems to be a
> rails bug, and the only deprecation warning we're
getting looks like
> being another case of Rails making an unwarranted
assumption -- we
> don't ever call the function it's complaining about.
>
>> Perhaps we should consider finishing up 4.1 first
and then start on
>> this?  When is Rails 1.2 due?
>
> What are we lacking for 4.1?
>
> And I'm not sure when 1.2 is due, but the noises at the
European Rails
> Conf seemed to to imply it was coming real soon now.
>
> -- 
> Piers Cawley <pdcawleybofh.org.uk>
> http://www.bofh.org.uk/
> _______________________________________________
> Typo-list mailing list
> Typo-listrubyforge.org
> http:
//rubyforge.org/mailman/listinfo/typo-list

_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

Matt Rose <mattrosefolkwolf.net> writes:

> What's the URL for it?

svn://typosphere.org/typo/experimental/restful

-- 
Piers Cawley <pdcawleybofh.org.uk>
http://www.bofh.org.uk/
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list

Matt Rose <mattrosefolkwolf.net> writes:

> On 28-Sep-06, at 12:17 PM, Piers Cawley wrote:
>
>> Matt Rose <mattrosefolkwolf.net> writes:
>>
>>> I say do it.  Every good project needs a good
ground-up rewrite  
>>> now and
>>> then, and I think typo is past due.  There's a
lot of cruft in  
>>> typo that
>>> just needs to go.  Can anyone say, AJAX Sidebar
activation?
>>
>> Actually, that's fixable. But it's been problematic
because it really
>> didn't sit well with the workings of the sidebar
object; we're
>> roundtripping to the database far too often when we
could/should be
>> letting the browser retain the information. I've
been planning a
>> rewrite for about a yonk though.
>>
>
> Is this why my new sidebar goes to the database to grab
the  
> information twice?
>
> I've been puzzling this one over for quite some time. 
I have a  
> sidebar that's pretty much complete now, however, it
takes well over  
> 30 seconds to load.   When I look at the logs, it's
because the  
> sidebar goes back to the database and loads the feeds
and builds the  
> feed, which is a pretty time consuming thing to do, and
does it twice.
>
> Why is this?

I dunno. I don't think it's related to the way sidebar
configuration
works though.

-- 
Piers Cawley <pdcawleybofh.org.uk>
http://www.bofh.org.uk/
_______________________________________________
Typo-list mailing list
Typo-listrubyforge.org
http:
//rubyforge.org/mailman/listinfo/typo-list