[pmwiki] Re: [Pmwiki-users] CVS features was: discussion on line breaks

[Bernhard.Weichel@t-online.de]

Patrick R. Michaud wrote:
> On Thu, Aug 14, 2003 at 01:42:47AM +0200,
> Bernhard.Weichel at t-online.de wrote:
>> Patrick R. Michaud wrote:
>>
>>>> I would really like to have a full title in addition to the title
>>>> which establishes the link.
>>>
>>> This gets a bit tricky in some of the other markup but is definitely
>>> doable.
>>>
>>
>> Isn't it just another field then. Or do you think to take the Title
>> from Markup and not from an extra field.
>
> Well, it depends on what you intend for the full title to be used for.
> It can't simply be a replacement for $Title/$Title_ in Fmt-variables,
> because lots of parts of PmWiki (and many customization modules)
> depend on those being the pagename.  That's what makes things a bit
> tricky.  So the question is, even if the pages store the full title
> somewhere, where/how
> will this information be used?
>
>> BTW: wikipedia has a checkbock which can indicate minor changes. And
>> minor changes are not recorded in the diff. In particular the most
>> recent diff
>> is removed and replaced by the minor change. Would that also be
>> possible
>> for PmWiki.
>
> Yes, it's possible to have PmWiki do something like this, but I'm a
> little concerned that the minor edits feature could lead to abuse by
> malicious edits, since it wouldn't be possible to restore a page to
> intermediate versions in a sequence of edits labeled as "minor"
> (which could in fact be major and/or malicious).

you are right.


> Failure-case scenario:  I add some major additions to an existing
> page, thus saving a new diff entry.  Another person makes a sequence
> of several "minor edits" to the page, perhaps to correct spelling or
> other items.
> As you mentioned above, the diff entry is replaced by new diffs
> containing the minor edits.  Now then, someone makes a malicious
> change to the page, but indicates it as a minor change, which also
> replaces the diff.  Because the diffs have been combined together,
> there's now no
> easy way to return to the version of the page after my major edit,
> or after the second author's minor corrections.  This is not to say
> this can't be overcome, but the algorithm for managing diffs has to be
> very carefully thought out.
>
> I'm also against combining the diffs associated minor changes because
> it then becomes difficult to track authorship of minor changes.  I
> think
> the real solution is not in changing how PmWiki stores diffs, but
> rather
> to keep all diffs (as PmWiki currently does), have a mechanism to tag
> certain changes as being minor, and have an option to combine minor
> diffs together when the page revisions is being displayed (in

This is an excellent idea and probably the way to go.

> contrast to
> trying to combine them when pages are being stored).  Alternatively,
> PmWiki could keep track of two sets of diffs--one set that tracks
> all changes, including minor ones, and a second set that tracks
> only the major changes.  Either way, this feature needs a lot of
> careful attention to detail.

>> I think this is also a very fine approach. Then we should have a way
>> to link regular WikiGroups with UserGroups in order to attach write
>> permission.
>
> As I mentioned in my listserv email, I'm not planning to add access
> control lists or page security features based on the usernames unless
> I hear a lot of clamor for it (and so far everyone is saying they're
> only interested in tracking edit changes and specifically not wanting
> user-based security).  However, since PmWiki's security features are
> a pluggable feature, it'd be possible for someone to redefine the
> security system to use ACLs and user-based security.

Yesterday, we had discussion about our new website strategy. And there
the hay the requirements:

 * some pages should be marked such that the can only be edited when
   being logged in. (priority A)

 * some groups should be edited only by a particular usergroup (priority B).

Can I be of any assistance in the current development of "Track Authorship",
e.g. writing docs, performing tests ...)?


regards Bernhard

--bernhard