Reduce module count

Assignment
Committee: 
Web Design
Assignment Date: 
2009-01-25T00:00:00
Assignee(s): 
Walt
Completion
Actual Completion Date: 
2009-03-23T00:00:00
Percent Done: 
12/31/1969
Expected Completion Date: 
02/26/2009
Percent Done Date: 
2009-03-23T00:00:00

Reduce the number of modules loaded to increase performance.

There are several categories of modules: essential functions (E), support features that we want to provide (F), deletable (D), present for testing debugging - go away at launch (T). Some modules have several submodules (n).

 Module  Cat.  Disposition  Date
 Masquerade  T  Delete 2/28  
 Adsense (3)
 E    
 CCK (22)
 E (19-20)
 Delete some if not used
 
 Core optional (17)  E    
 Throttle
   may need to install for performance under load
 
 Node Import  T  Delete 2/28 (reinstall brieflywhen needed)
 
 Table API
 E  part of store
 
 Event (3)
 E    
 Image (5)
 E (2)
 alternates available
 
 Imagecache (4)
 E    
 JS tools (4)
 E    
 fquery  D  don't know if it is needed (probably not)
 1/27
 Location/Gmap (6)
 D  Depends on which Google map solution we choose
 1/28
 Mail (6)
 F  some overlap with other modules
 
 Messaging (6)
 F  SMS messaging - may not support
 
 MySite (2)
 F Not on now - customization per user
 
 NYNJ Custom
 E  Trellon theme support
 
 Notifiacation (7)
 F  some overlap with other modules
 
 Organic Groups (4)
 E    
 Actions
 F  may need to add  
 Addthis  F  slows performace sometimes - touches external site
 
 Admin links
 T  Delete 2/28 -site building aid
 
 Bio  F  Profile addon
 
 CitizenSpeak  F  CapWiz replacement??
 
 Clone (2)
 T  Delete 2/28 and reinstall if needed heavily
 
 Colophon  T  Delete 2/28
 
 Contact Forms
 F  not currently used??
 
 Date API
 E  CCK date fields
 
 Diff  F  revision helper
 
 Find URL alias
 T  may want to leave  
 Form (2)
 E  AJAX submit  
 FAQ  F  can be done by hand
 
 Glossary  F  may omit feature
 
 Google Analytics  E    
 Helptip  D  may omit feature  1/27
 Image caption
 E  not using yet
 
 Image picker
 F  one solution
 
 IMCE  F  one solution
 
 Job posting  F    
 Lightbox  F  not working because js_update fails
 
 Logintobogan  F  extends login process to allow either uid or email  
 Mailhandler  F  accept updates via email - not working yet
 
 Metatags  E  SEO  
 Mollon  E  Spam prevention
 
 Nice menus
 E  part of site design
 
 Node browser
 F  don't know how to use - good candidate for delete
 
 OpenID  F  extends logon - to universe
 
 Page Title
 F  don't know if used - good candidate for delete  
 Pathauto  E  SEO node names
 
 Poormanscron  T  Delete - have real cron 1/26  
 Printer friendly
 F  off - may need
 
 Profile privacy
 F  probable - depends on what we collect  
 Quick tabs
 F  don't know if used - good candidate for delete  
 Related links
 F  used - discuss deletion  
 Role change notify
 F  could do manually
 
 Search engine ref
 F  better statistics
 
 Secure pages
 E  by store
 
 Signup  E  event registration
 
 Site documentation
 T  Delete 2/28
 
 TinyMCE  E  rich text editing - look into newer version and compression  
 TinyMCEconfig  T  I think a one shot deal setup helper
 
 Token  E  required by many other E  
 Top Searches  F    
 Tracker 2
 F  replacement for core Tracker
 
 uBrowser  E  by store
 
 Update status
 E  required for site maintenance  
 Username AJAX  F  nicer register helper
 
 Web links (2)
 F  Replicates Links page from old site (auto not found help)
 
 Webform (2)
 E  easy generate forms
 
 XML sitemap
 E  off currently - needed for SEO
 
 Panels (8)
 E  site design
 
 Relevant content  F  not useful until we do more with taxonomy
 
 SWF tools  E  Think this is what plays the flash headers (if not delete)
 
 http:BL  E  spam control
 
 Subscriptions (8)
 F    
 Template (2)
 ?  Don't think we use
 
 Ubercart (many)
 E  only features we need loaded
 
 Views (10)
 E    
 FiveStar (2)
 F    
 Workflow-NG  E  by store
 

Deleted a bunch that I know we do not use and suspect we won't use.

Comment: Please be relevant, civil, non-commercial.

egoodell's picture

There are 80 modules in the list.  Trellon recommended reducing by 6, to approximately 13 modules, yet 32 are listed as "Essential" and another 36 as desirable "Features." 

So, there is a huge gap between the recommendation and current practice.  We are going to have to be more agressive. 

Trellon seems to be suggesting eliminating modules in favor or programming features into the core code.  We should investigate cost, find the money and move forward.  In the meantime, we should do without some of the features.

webmaster's picture

There are more the 13 pieces to Ubercart alone. The factor of 6 is completely unrealistic. We need a realistic estimate based on other sites that do e-commerce as a minimum starting point. Drupal has a lot of the properties of protecting environments where one finds out that they are all connected and you can change anything without effecting everything. For example, we could do everything that Panels does with regions, which are part of core, but panels are also part of Organic Groups where they cannot be replaced with regions without reimplementing OG. There are other approaches which might work. The site can be split into multiple subdomain sites with shared user tables. So for example the store and Groups could be separated off. Trellon recommendations on this might be useful without doing a lot of experiments to see how it goes. I don’t think it reduces the module count for Groups very much so it would still be big. It is also inline with a projected direction of using something like CiviCRM for our membership management database, which is its own huge pile of modules.