Last modified: 2006-01-16 04:41:35 UTC
The software should add some info about the edit when someone doesn't fill out
the edit summary field, making patrolling easier on everyone. For instance, if I
make a minor change to an article, changing "5KV" to "5 kV", I just press alt+s
out of habit, without filling out the summary. What would be ideal is if the
software just autogenerated a summary of:
16:15 Electricity (diff; hist) . . Omegatron (Talk) (→Voltage - "5KV" → "5 kV")
In cases of vandalism, this clearly labels it for those patrolling recent
changes or watchlists:
16:15 Politics (diff; hist) . . 220.127.116.11 (Talk) (→Political concerns - "is a
member of the" → "IS GAY!!!")
Since the autogenerated summary is colored, it can be trusted and can prevent
the need to waste bandwidth, time, and server load from checking a lot of diffs
by anon editors that are simple spelling fixes and such.
"But we already have a policy of encouraging edit summaries! They should be
filled in!" So what? A lot of people, especially vandals, simply don't fill
them out. It would be tremendously helpful if the software generated a crude
summary itself in these cases.
This saves precious server resources by processing a diff only once, preventing
the need for many people to load that diff again in the future.
More details and examples at the included URL.
(In reply to comment #0)
> The software should add some info about the edit when someone doesn't fill out
> the edit summary field....
This is a similar to bugzilla http://bugzilla.wikimedia.org/show_bug.cgi?id=1307
"Auto-Edit Summary: if few changes ("smaller" than a threshold) have been made,
automatically create an Edit Summary from diff results"
This is the germ of a good idea. I might not like to see the mini-diff right on
top of or sharing the edit summary field; I'd rather have another column in the
Implementation might be very tricky. The diff mechanism already tends to choke
and get confused, failing to equate large blocks of identical text. I do agree
that a mini-diff of some sort would prove extremely useful.
I bet a user script could be written based on this diff algorithm:
I started to look at it but I am not that good with js yet.
I'd go for something with the name of the section that was edited, if a section
*was* edited, then the change in size, in bytes, e.g.:
→ Voltage (+1)
*** This bug has been marked as a duplicate of 1307 ***