[dev] Rationale for some fields in Mnemo, Turba and Nag

[Chuck Hagenbuch]

Quoting Luciano Ramalho <luciano at ramalho.org>:

> I've noticed the memo_owner is not necessarily a username, but is  
> the id of a share (which may be a username by coincidence). Am I  
> right?

Exactly.

> I'd like to know why the memo_uid field is needed, what is it used for?

It's used for syncing, things like vCards, iCalendar attachments, etc.  
The UID remains the same on multiple systems (another calendar server,  
a PDA). It's not needed for data that isn't synced.

> And finally, I've noticed that in more recent sample apps such as  
> Mopple and Properties, you're adopting the practice of using a PEAR  
> sequence to generate the id, instead of the md5(uniqid(mt_rand(),  
> true)) method used before. Is that an emerging coding standard in  
> Horde?

Only Turba originally used the big random string method. The other  
apps were changed to it when we first started dealing with UIDs, but  
that turned out to have to be a seperate field. They just haven't been  
turned back yet.

-chuck

-- 
"we are plastered to the windshield of the bus that is time." - Chris