Re: [pLog-General] developer tree

[Oscar R. <os...@re...>]

> In general, what has changed in the devel tree since the 0627 release?
>  Are there any concerns you have about it?

Well, not much has changed... Unfortunately I haven't had much time to  
work on anything. I had some time to develop the code to make the  
wizard be able to upload from 0.3.x and that's pretty much about it.

However the main concern at the moment is that the search engine does  
not work with multibyte character encodings such as utf-8. The problem  
is that mysql 4.0 only supports iso-8859-1 as the character set in  
FULLTEXT indexes (this kind of indexes are used by the current search  
engine implementation) but mysql 4.1 already offers support for utf-8.  
However, my idea was not to depend on mysql's implementation *too  
much*...

There is a taiwanese cluster of users who told me that they would have  
a look at the code of MediaWiki (the wiki tool used in wikipedia.org)  
because it implements a utf-8 based search engine on top of mysql  
FULLTEXT indexes, and it works even if we're only running mysql 3.x.  
I've also had a look at the code myself and does not look that  
complicated but I'm still considering how to proceed. Not sure if we  
should tell everybody that mysql 4.1is needed or try to spend some time  
working on reimplementing the search features... If you keep in mind  
that the chinese/thaiwanese community is the biggest one we have, it  
wouldn't be too nice to leave them out of such interesting feature.

Don't know, I'm open to suggestions...

> I noticed that two of the big sellers plugin-wise haven't yet been
> made:  the template editor and the advanced search plugin.  What's the
> status on these?  I may be able to help with the template editor,
> though I haven't a clear idea on how it would work -- would there be a
> separate /template/blogId/ folder for each blog?  Seems like a waste
> of space to me, especially for those blog community "providers" like
> balearweb and lamevaweb.  This is what I am thinking, though I'm not
> sure if this is a better process or not.

I thought I had found somebody some time ago to develop the template  
editor but I haven't heard from him in a while... Very roughly, the  
idea was to allow users to create a copy of a global template and then  
save it as a blog template so that they could work on it. They would be  
allowed to create new files, edit files, access to all the smarty code  
and everything but I guess that this is something that not many users  
would be comfortable with...

I will check with him to know what the status is, I'm also very  
interested in such functionality!

> Template authors should submit a general template structure with
> default styles.  Additionally, authors must include with it a profile
> storing all the fields for the different elements within the template.
> The template editor uses its own table , one row per blog with just
> three elements:  blogId, general template structure name, itemized
> array of the colors/font styles.  The editor wouldn't allow users to
> actually modify templates, but would let them change text decorations,
> sizes, borders, and such.  This still needs some working out to do.
> The templates could then instead implement a command with
>
> <link rel="stylesheet" type="text/css"
> href="/css.php?blog={$blog->getId()}" media="screen" />
>
> where css.php would generate a stylesheet based on data from the table
> applied to the stylesheet template (also new, create a new required
> template article, "stylesheet.template" with listing for all the
> classes but instead of values replaced with something like {
> $templateeditor->styleSheet($blogId)- 
> >getClass("content")["key_name_like_p_.img"]
> }

The thing is, if we only allow users to change the stylesheet, why  
don't we just store a .css file somewhere in the disk and allow users  
to edit the file directly, rather than doing complicated stuff such as  
storing all the different css classes in the database and so on? The  
idea would be exactly the same thing that WP is doing with the "style  
switcher": the same template for everybody but allow users to use a  
different stylsheet

> I have no experience with writing efficient SQL queries, and really
> couldn't begin to help with writing a search engine plugin.  Aside
> from that, I could help supply a new design for the admin panel and
> possibly help with a "community plugin" that's being debated right now
> in the forums.

I asked the user 'blacksnday' from the forums to be more specific  
regarding these "community" features. They sound quite interesting but  
I don't have the time to take care of them... Would be great if you  
could! :-) I believe that these things can be integrated as plugins  
unless they require some deeper ties with the core and if that's even  
the case, we could discuss whether to allow this into the core or not.

> Oh yeah, if you have time check out what I've done to my personal
> website (bliang.com).  I don't know what you guys think of this, but
> I've quietly been working with the plog development tree to create a
> non-bLog pLog based website.  What do you think?

Wow, that looks really cool!!! It would be interesting if you could  
write some documentation about how you did it and so on, in the wiki,  
so that potential users interested in using plog as an engine rather  
than a tool could follow your same steps...

Oscar