Fwd: Re: [pmwiki-users] Is there a way to tell if anyone is editing a wiki page?

Joachim Durchholz jo at durchholz.org
Fri Apr 15 15:43:06 CDT 2005

Radu wrote:

> At 01:27 PM 4/15/2005, Joachim Durchholz wrote:
>> What you don't see is whether the user has aborted the edit session.
>> One scenario is when the user presses the Back button and the browser 
>> decides to display the page from its cache - the server isn't notified 
>> of that even.
>> The other is when the user simply closes the browser window. Again, 
>> there is no feedback to the server, so PmWiki doesn't even have a 
>> chance to find out.
> Yes, however, when the wiki authors are conscientious(sp?) designers, 
> this kind of behavior is much less likely to happen.

Wikis live from the idea of placing an as small as possible set of 
restrictions and obligations on their visitors. Having to be careful 
definitely is beyond that.

(Heck, PmWiki doesn't even have a "Change summary" field.)

 > And adding a
> timeout for the 'notelock' I was suggesting would help with that too. 

Timeouts tend to confuse the users. PmWiki will behave differently 
depending on whether it's more or less than 30 minutes that I needed to 
do that edit... yuck.

> Anyway, a project member that notices a page has been note-locked for 
> too long could call/email the member who 'locked' it and ask what's 
> taking so long.

Well, right, but my experience tells me that this is always an 
annoyance. Also, this just encourages members to invent plausible 
excuses. Topic hogging becomes a problem. Pessimistic locking encourages 
people to lock their topics as long as possible.

Optimistic locking, on the other hand, does the exact opposite. The 
longer you take to finish your edit, the more likely somebody else 
finished his edit, and you have to incorporate his changes into yours. 
That's work, we're all lazy, so people try to keep their editing times 
as short as possible.

(I have lived under both kinds of locking regimes. Pessimistic locking 
is simply a royal PITA.)

>> So if you want to find out whether a page is open for editing: not 
>> possible.
>> However, this isn't a serious problem. If two people edit the same 
>> page and save, the second one to save will be notified of the conflict 
>> and get a chance to integrate his changes with those of his competitor.
> Many times in the little while I ran this wiki, I noticed people saving 
> pages without noticing [all of] the merge brackets. It gets ugly, lemme 
> tell ya.

Now that's something worthy of consideration. Maybe PmWiki should refuse 
to save a page that still has merge brackets in it? (On the page that 
refuses the page, it might present a "Save anyway" button so that people 
who are just too confused to get the merge done can still save their work.)

>> (This strategy is generally called "optimistic locking" in the code 
>> management literature.
> Cool. but wiki users are not machines. They can [better] solve different 
> sets of problems than a cvs or rds system can.

Not sure what you mean with that. CVS conflicts aren't machine-created 
problems presented to humans - they are humanly-created conflicts, 
detected by the machine, and delegated back to humanity for solving.

This is just the same as with wiki edit conflicts.


More information about the pmwiki-users mailing list