Re: [Linkbat-devel] Linkbat Dataflow

[James M. <lin...@ji...>]

Hi All!

Sorry, I'm doing it again. I forget to keep a discussion with Hal Gottfried on 
the list. Hal and I have been discussing the XML implementation in the data 
files and he has said he can write the DTD when we get the model finalized. 

So, to sum up our discussions, we both seem to agree that accessing the data 
will be speeded up by a conversion to either CSV or a database. Hal made a 
couple of comments that me think that we could actually create a lot of the 
pages in advanced and not at run-time, thus speeding up things even more so. 
Since the data is static, once copied to the server, there is no reason to 
parse most of it on the fly.  For example:

- Content pages. They are parsed at runtime to create the links and popups for 
the glossary, links to other tutorial pages, etc. 
- Glossary pages. Created completely at run time and include links to other 
glossary terms and pages that references this glossary term. 
- MoreInfo pages. Created completely at run time and include links to other 
sites. 

I see no reason why these cannot be created in advance. We would just need to 
make sure that whatever display mechanism we use knows how to load the 
correct page. To me that's a heck of a lot easier than creating the pages on 
the fly. 

This brings up a **huge** question, which is particularly of interest to 
Shanta. If we create all of the pages in advance, there does to seem to be a 
need a pressing need for an SQL Database. The reason for it was the 
efficiency during data access. So, the question is whether we gain anything 
by creating one. In my mind, it doesn't matter if the pre-processing takes 
longer as it would with an SQL DB.

How do the rest of you see this?

Hal suggested including the PageID as an attriube to the <Type> tag and we 
were both thinking that we could the sub-type as an attribute of the <Type> 
tag.

<Type id="45">Page</Type>
<Type url="http://www.linux.org">MoreInfo</Type>
<Type subtype="verb">Glossary</Type>

This makes sense as the sub-type is a characteristic of the type and not the 
actual KU.

For the questions, we have the issue of an empty <Reason> tag as a reminder to 
add something. It might be simpler to have the text of the Reason an 
attribute of the tag, like this:

<reason answer="why this is correct" />

If it's empty we don't carry arounf the extra baggage.

Did I forget anything important, Hal?

Regards,

jimmo

-- 
---------------------------------------
"Be more concerned with your character than with your reputation. Your
character is what you really are while your reputation is merely what others
think you are." -- John Wooden
---------------------------------------
Be sure to visit the Linux Tutorial:  http://www.linux-tutorial.info