Most "recent changes" are caused by those who like to correct grammar and change formatting. This is generally not at all interesting as a "Recent Change". For those who like doing such changes I have a humble suggestion.
RT 14Nov2008 - still virtually worthless. To all the edit Gnomes: Thanks for nothing
CMcC FWIW, I find the edit gnomes' contributions worthwhile. I'm glad they take the time to improve the wiki's contents. I object to RT's denigration of the gnomic contributions, to his mischaracterisation of the Recent Changes' worth, and to his disruptive formatting (bunging a one-liner at a random spot in a page, then complaining that people want to reformat?)
I'll cheerfully agree that Recent Changes isn't perfect, but I don't think it has to be perfect to be useful. There are constraints (developer time and priorities, requirements of security and protection against abuse, communal unwillingness to require login for edit, my personal lack of caring about the 'problem', or creative solutions to it) which keep this problem in its current state. If RT wants to put the effort into actually coding a creative total solution, I'm happy to read his or her code.
RT Gosh, how persuasive. "I object..(!)" Wow. I have "disrupted" the formatting of this page! Really. Hallucinate much? Then you trot out the fallacy that I'm demanding perfection. That I don't understand "there are constraints". What wisdom. Gee, I feel so silly now. As surprising as this concept may be, I actually made a practical suggestion when I created the page. Other commenters also discussed practical possibilities. In the end nothing has been done and nothing has improved. More than half the pages marked as "recent changes" to me seem to have literally no cognitive change at all; the red and green sections read identically. These so-called changes are simply information pollution caused by apparently compulsive editing behavior. Alas, the offer of your happiness for my code is declined. Happy bunging :*)
CMcC You appear to be simply restating the case, ignoring the constraints, and being quite rude. I checked your contributed code base before suggesting it, and I didn't really expect you to contribute working code, but if you want changes 'in the end', that's the only way you're likely to get them.
It's probably worth restating that there's no shortage of commentators, but the constraints listed include developer time. I don't understand how you can maintain that these changes are so important that someone else needs to work on them. If they are really that important to you, then by all means work toward making them. If they aren't, then why expect someone else to work on them? You're not a commentator, you're a promoter.
How about all those formatting / grammar gnomes use some coming login name prefix/suffix.
For example, Instead of LV ==> copyfixLV Instead of dkf ==> copyfixdkf
Make up your own tags but make them obvious. Thanks for you attention, RT
SMH 20081115 just wondering of a copyfix change could effectively hide a significant change...
JH Other wikis have a checkbox for "minor change" - we should possibly adapt that for this wiki.
anoved I concur. The wiki maintenance work is important, but I would find the Recent Changes page most useful as an overview of substantive new content or discussion.
DKF: A "minor changes only" flag might be nice. But the other idea (special names for when gnoming) is harder because I may actually switch back and forth a few times when doing it.
TP Replying to DKF: yes, maybe harder for the wikignomes, but any real content changes the wikignomes make get lumped into my wikignome visual filter, and thus ignored. I also have to comment that the recent Category formatting trend is simply annoying.
AMG: If we go the "minor change" , I suggest also having a checkbox (or link or whatever) on Recent Changes to show everything, including minor changes. Wikignomes might like to see that. :^)
Twylite: I'm just starting to watch the recent changes via RSS and I've noticed some possible problems (or areas for improvement):
Sly I suggest that some metric of change amount should be displayed. E.g. added+deleted+modified characters/lines.
This should be done on both 'Recent Changes' and '_history' of pages.
Make the minor change or classification a mandatory field. In many wiki's it is optional with a certain default value and then it can go wrong either way (minor change registered as major and vice versa). By giving it no default value and make it mandatory you get the 'right' type of change (if filled in truthfully of course) -- Jar.
DKF: I'm worried about abuse, e.g., wikispammers putting bad links in as minor changes. And in any case, most people don't work as wikignomes.
anoved: What if minor changes were simply marked in the recent changes list with an icon or a different style? That way, all modifications would still be listed, but people interested only in new content could easily see which changes they can ignore.
MG One potential problem with that is that if a page changes several times in one day, only the most recent change shows up. So someone could add a huge amount of content to a page, then someone makes a minor change to it to fix a spelling/formatting error, and noone notices the real change...
anoved: Good point. What if each recent change item was followed by a series of little minor/major icons? These could take the place of the single diff icon we have now to show the diff for each successive change. Major diffs could be represented by the solid delta triangle icon used now, and minor diffs by a hollow 'outline' triangle. I suppose a separate 'all of today's diffs' icon or link would be useful, too. These arrangement would help the recent changes page show popular or contentious topics as well as recent changes, be they major or minor.
SupremeXeN.com KommandEr : Title of page is too inflammatory. Recent Changes in its current incarnation is how this page has attracted discussion. Tcler's Wiki does not get more than 50 changes per day, on average (statistics derived from observation). Things are very different on this wiki than on others. Aside: Documentation and chatter on the same page leads to difficulty for newcomers.
Sly: 'Minor change' checkbox is also 'virtually worthless' :) There's no objective measure of 'minority' of a change. The reader might have another opinion about significance of page change than writer. I never liked the 'minor change' option of other wikis. Any of objective change metrics would be more suitable here.
RLH The original rant was "Most recent changes are caused by those who like to correct grammar and change formatting." So it wouldn't be worthless to the person making those changes so they never show up on the "recent" page now would it?
Zarutian: Sly, so you would rather recommend an slider with values between 0 and 10 that signifies the editor's subjective measure of 'minority' of a change?
KJN: there are not many changes per day, as SupremeXeN.com KommandEr has noted. [non Recent Changes] could show the last change of each page, and if that is a minor change, also show the last major change (so some pages will appear twice on the page); the "minor changes" could be flagged so people will not follow the links if they are hoping for new content. 'Minority' is decided by the author: most of us will trust the Wiki Gnomes to set this flag correctly; "minor change" set by an unfamiliar person will attract attention, so it will be hard to hide abuse.
LV: How much annoyance are wiki updaters willing to endure for the sake of the wiki? I mean, what if to edit a page you had to supply an id - that id field being filled in by default by the login value? Then, if the id was close to the button to submit the change, one could consider whether the special login approach, if we didn't want to specify minor edit type things.
Another option, I suppose, would be to just let the quality of the wiki erode by not making changes to point to updated URLs, fix spellings, etc. I don't think that is the point of the page's rant.
I suspect the page was created after someone had a bit of free time and went about fixing some pet peeve. So a string of minor edit updates occurred.
Here's the thing about a minor edit checkbox. Sometimes those things are forgotten. I know I keep forgetting to tick it on Wikipedia when I'm editing there.
And while some conventions might seem obvious (if you add a category or fix a typo only, then that is a minor edit), what if someone fixes a single line of code that was wrong? It might be a single character fix ... is that a minor or a major edit? On the other hand, I've seen long diatribes that I would consider a minor edit because it added nothing useful to the context.
I used to watch news headlines at CNET[L1 ] or one of those sort of tech type web sites. Authors would tweak their articles, sometimes correcting typos, sometimes correcting facts, etc. When such changes were undocumented, it was always frustrating to me, because I had to reread the whole page and hope that I could remember what had changed. So I understand the frustration someone who stops by and sees a long list of pages. One thing that is nice though is the delta on the recent changes page - a simple click (at least in most web browsers) can quickly show the page positioned near the first of the changes on the page. That is something nice from a site browser. And if RSS supports pointing to the diffs, that would be a nice enhancement ... maybe someone more familiar with RSS could submit an enhancement request at the wiki's project page.
I've thought for some time that a move from content level pseudo tagging to some other kinds of tagging might prove useful. One thing I've wondered is whether there were a way to replace the category links with software that would generate a tag cloud based on the contents of the page, along with some sort of out of band additional indexing terms or something. Alas, I've not had time to experiment with the ideas. P.S. Colin, sometimes I do add content...
escargo 07 Jan 2008 - I had filed an enhancement request that in addition to the delta there would be some other thing to click on to get you all the changes made in the last day. (In fact, if there were room, you could have a list of numbers, like 1 2 3 7 30 to get the changes for the last few days, week, and month.) However, since you can do this manually by going to the page and clicking history, and then picking the date yourself, this was not seen as a useful enhancement. I think it it would be so much more convenient this way.
EKB I've been trying to pay attention to how I use the Recent Changes page, since the frequent changes weren't actually bothering me. I noticed that I:
This pattern-matching works pretty well for me, but it could be partly automated. That is, instead of asking people to check off a "minor edits" box, have the wiki look for a long sequence of edited pages by the same person, and put those in gray/italic/something so it's easy to scan. It would certainly not be perfect -- it would give some false positives & false negatives -- but if the action taken isn't drastic, then it may be more a help than a hindrance. For what it's worth, I do like to see the wikignomed pages listed on recent changes, because sometimes a page is brought to the top that hasn't seen activity in a while, and I say "Oh, yeah, that was a good one!" and I go and remind myself of its contents.
MG I do much the same as EKB, except I don't bother to scan the pages where several have been edited by the same person in quick succession any more, as they're almost always all minor changes (formatting, spelling, etc), unless it's an old page resurfacing that sounds interesting. (Unfortunately, I probably do miss a major update from time to time because of that.) I think a minor edit checkbox would be good, though, that just made the page be displayed in a different style (paler, italic, something). Perhaps Recent Changes could show the page's last major update for that day, if it's had one, or its last minor update if it hasn't had a major one today? That way, minor updates wouldn't cause extra spam, or make people miss a major update that happened a few seconds earlier, but wouldn't be totally overlooked either.
CMcC 08Jan08. We discussed the 'minor edits' endorsement at the time we were re-making the wiki. I held out against it (over quite persuasive arguments) because we were in the middle of an unperson-attack which had brought the wiki down. Some of the first long-gnoming sessions were in fact undoing his damage. The problem I have with 'minor edits' is that unperson uses the technique to hide major vandalism. The 'minor edits' facility entails trust - you would have to trust every person and every unperson on the wiki not to lie about the changes. I thought that the only way around that was to institutionalize trust in wikignomes, which would both change the structure of the society (it's a kind of privilege, see) and require some sort of authentication, which would be a serious departure from tcl wiki norms, and be more work than I really wanted to do.
In the meantime, I use the algorithm outlined above - ignoring runs of changes by the same person. And I am very grateful that LV, dkf and others spend some time gardening the wiki, improving the quality of the soil. I actually think that my inconvenience, and bearing it stoically, is part of my contribution to their efforts.
If people want to boot around ideas to reduce the entropy of Recent Changes, that's great. If something definitive and original comes out, that's great. If a consensus arises, that's great. If someone wants to provide patches which encapsulate all of those, that's the greatest of all.
Lars H: The unperson argument is somewhat diluted by the fact that clearing a page is always treated as a sort of minor change: hidden away on a separate page, linked to without any indication of whether new items has appeared or not. While the particular unperson that haunted us typically did the opposite of clearing pages, future vandals may follow other patterns.
As for the Recent Changes page, I strongly support escargo's suggestion about direct links to diffs for the last day etc. Omitting those that would give the same result as the previous link would even make this a valuable indication about how active a particular page is (but it is of course easier to just make them fixed #diff1, #diff2, #diff3, #diff7, etc. links like the current #diff0 and not figure out what that means until receiving a request).
MSH 08 Jan 2008 I also support escargo's suggestion as I often remark that several large chunks of modifications (as here on this page today) are masked by one last small change to spelling at the end of the afternoon, I often wonder what useful bits I'm missing not re-reading the whole page just to see what else may have changed today or this week. on smaller pages it is often obvious that something has changed but larger pages are much more difficult, and a shortcut to the history/date selection would be extremely useful.
LV I'm trying to better understand what is being requested. Right now, on the Recent Changes page, one has the link to the page, the login id (most times), the IP address, and a link to a diff of the change they made. If one visits the page itself, and clicks on the history link, one then has the ability to see a plethora of ways of looking at the changes made over the history of the page.
Is the above discussion asking for a reflection of the history page out to the recent changes page? Or some subset of changes to appear in the RSS feed? What if there are two dozen changes in a day - something not out of the question, on a page like this, or whatever the recent question page is, etc. Are you wanting just the most recent N links, or as many as are necessary for the last 24 hrs? The clearer we can get about what is desired, perhaps even with someone mocking up what the page would look up, then the more likely someone might try to hack something up.
However, I do strongly suggest that if it hasn't been submitted to the google project page for the wiki, that someone write it up once the brainstorming settles down. That way at least the idea is there for developers to think about.
Lars H: This request is primarily for quick links (from Recent Changes) to differences based on (relative) time rather than revision number. Secondarily I observed that existence of such links could be used as an indication of how long the current modification activity has gone on (although that would have the effect that the history has to be consulted when generating the Recent Changes page, which could be undesirable).
[L2 ] is already a report of the issue that getting the changes for the last day is too roundabout, but has (as escargo clearly states above) so far only received the reply that the information can be obtained — with three times the effort! The idea was also added to WubWikit Problems, back when that was where issues were reported.