PhpDig.net - View Single Post

circusnews · 05-29-2004, 08:13 AM

Nope, no project for it yet - but feel free to start one. At this point I am thinking that this is going to be a %50 rewrite of PHPDig to get it to work. And if I am going to go that far, I may as well make the whole thing work better.

Here are my thoughts on this:

- Database Abstraction

- APIs

- Seperate out PHPDig in to Nuke mods:

-- Crawler

-- Search Engine

-- Search interface/API

===========================================
Abstraction Layer.
One of the things that bugs me about PHPDig is that it only works with MySQL. If we are porting this to Nuke, lets take advantage of the database abstraction layer in Nuke.

===========================================
API Abstraction Layer.
The more I think about it, the more I don't like the idea of trying to map PHPDig VARS to Nuke equivs. It's messy and problematic. Given the modular nature of the PHPDig code, if we are going to do this, lets do a rewrite with Nuke API's, and then rebuild a standalone option as well.

===========================================
Modules.
The more I think about this, the more I think that it makes good design sence to have the crawler and search engine as seperate modules. By doing this, we also make it possible for others to develop additional add-ons to this system that we have not yet thought of.

===========================================
Modules - The Crawler
PHPNuke has a lot of modules that would seem to act as natural starting places for the crawler to start (web links, news story links, etc), and more modules that are developed all the time. By seperating out the crawler, and providing a place to plug in, we can have the crawler automaticaly crawl these things (as set by admin)

===========================================
Modules - The Search Engine
The search engine itself should be as modular as possible. Thankfuly, PHPDig is very modular.

As we do this, lets remember that PHPNuke already has a seach engine used for the site itself. While well intentioned, I hate it. It does however, have a lot of things we can learn from it

1. Searching the existing tables, and linking to the proper module search feature. While a big part of me does not like how this works, its a lot better than having PHPDig reindex the whole site. What I would really like is a method of allowing you to choose (via drop down menu) what index you want to search. I know it will get more complex than that, but so be it.

2. Off site searching. Lets continue to make it real easy for other site to include a seach bar.

Other thoughts?

05-29-2004, 08:13 AM	#12
circusnews Green Mole Join Date: May 2004 Posts: 8	Nope, no project for it yet - but feel free to start one. At this point I am thinking that this is going to be a %50 rewrite of PHPDig to get it to work. And if I am going to go that far, I may as well make the whole thing work better. Here are my thoughts on this: - Database Abstraction - APIs - Seperate out PHPDig in to Nuke mods: -- Crawler -- Search Engine -- Search interface/API =========================================== Abstraction Layer. One of the things that bugs me about PHPDig is that it only works with MySQL. If we are porting this to Nuke, lets take advantage of the database abstraction layer in Nuke. =========================================== API Abstraction Layer. The more I think about it, the more I don't like the idea of trying to map PHPDig VARS to Nuke equivs. It's messy and problematic. Given the modular nature of the PHPDig code, if we are going to do this, lets do a rewrite with Nuke API's, and then rebuild a standalone option as well. =========================================== Modules. The more I think about this, the more I think that it makes good design sence to have the crawler and search engine as seperate modules. By doing this, we also make it possible for others to develop additional add-ons to this system that we have not yet thought of. =========================================== Modules - The Crawler PHPNuke has a lot of modules that would seem to act as natural starting places for the crawler to start (web links, news story links, etc), and more modules that are developed all the time. By seperating out the crawler, and providing a place to plug in, we can have the crawler automaticaly crawl these things (as set by admin) =========================================== Modules - The Search Engine The search engine itself should be as modular as possible. Thankfuly, PHPDig is very modular. As we do this, lets remember that PHPNuke already has a seach engine used for the site itself. While well intentioned, I hate it. It does however, have a lot of things we can learn from it 1. Searching the existing tables, and linking to the proper module search feature. While a big part of me does not like how this works, its a lot better than having PHPDig reindex the whole site. What I would really like is a method of allowing you to choose (via drop down menu) what index you want to search. I know it will get more complex than that, but so be it. 2. Off site searching. Lets continue to make it real easy for other site to include a seach bar. Other thoughts?