Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
In a parallel discussion
http://www.math.ntnu.no/~stacey/Mathforge/nForum/comments.php?DiscussionID=1838
we discuss interesting Urs's idea about creating some sort of tags for pages or sections which show verification of factual content of the page (not the graphical quality or finished status).
We have opinions about what we wrote, like I am happy with the content, and we know what we have done and what we have not, if something is under construction, or I feel content with it. Sometimes we don't, sometimes there is consensus, sometimes not. Part of the sentiment is channeled via emails, nForum discussions and so on. We can try to extract the information in some more efficient form, or to find better ways to archive or find such information or to supplement in future discussions this information with more structured forms of information.
In this general sense I see, in addition to the approval tags which other discussion is more concerned with, also need or potential in more wide context of getting to the side information. Andrew warned us many time that the nForum is not meant to be of permanent nature, indeed our discussions here are even less polished than nLab, less structured in their organization and often not interesting after a while. On the other hand, wikipedia has talk pages where the issues about the state of some web page are discussed (like disagreements among contributors, and explanations for possibly misunderstood actions and edits). LaTeX has comments lines with percentage marks, similar deviced exist in most programming languages. Web and cweb of Knuth have roles in documenting "literate programming" and nlab has the query boxes which have not only the usage for creators but in the cases of unsettled state of some pages as information for other users who came across the page. There was a problem with accumulating old query boxes and there was a fact that while some pages were about content others were just fo technical nature and hence more akin to the need which is covered by talk pages in wikipedia.
Finally, while nForum has a search button and nForum is also google searchable and is meant to be temporary I see much of possible benefit in being able to see discussions directly refering to a particularly given nLab page (in various logically interesting orders, including by time of writing the entry containing the link and by time of the last active discussion in the same thread; or sorted by latest changes versus other subsections of nForum). For example, I may be interested in motivation for some page which may be hinted in the entry of the creator which is maybe recorded in the entry of nForum.
So what do people think about above wider issues, and in particular about
1) having some sort of talk pages attached optionally to particular nLab pages (my idea was hosted at nForum but with back link at that entry in nLab, but others may have other ideas)
2) having some sort of nLab-page specific search for particular page mention in nForum, which would be accessible by a single click from a nLab page, without typing the name of the nLab page and including all the aliases ? Would that be useful to others as well and how difficult is this to implement ?
I have several times wished for a way to put “comments” into the nLab, like % lines in TeX. Having a single “talk page” doesn’t quite serve this purpose, because there could be a lot of talk about a given page, and a comment is likely to apply specifically to a particular location on the page. But of course the nForum lets us have discussions on any subject, including a particular piece of one page, so all that would be needed for that is a way to associate the two. Perhaps there could be some CSS thing one could insert on an nLab page which would display as, say, a small unobtrusive “Talk” link on the right-hand margin taking one to a specified nForum discussion. Perhaps it could also simultaneously create an #Anchor on the nLab page, so that the nForum discussion could easily link back to the specific page location.
I am very much in favour of having appropriate discussions about a page in a single place. But I would like that place to be the page itself, not a separate talk page, at least while the discussion is active. This is much easier to see. A separate page as an archive of important discussions? (which is not every discussion).
This would be very handy. Just something to search for [[name]] (including redirects) in the sources of comments would be great. We could put an automatic ‘Search for discussions of this page on the nForum.’ link on every page. A fancier version could include the latest date of any such comment (although I don’t think that this would actually help me).
Invisible comments in the vein of TeX’s % or HTML’s <!-- --> can also be handy for very short notes to potential editors; Wikipedia uses this, for example, to remind editors that some issue was discussed and decided in the past. But I think that this is a separate idea from Zoran’s (1) and (2).
As a ’quick fix’ to see if something along these lines is (i) useful, (ii) easy to maintain etc. and so that we get experience on how a more permanent solution might work, why not use query boxes as discussion boxes. Start a query box with DISCUSSION, of COMMENTS, as a title (or if the discussion takes off, ONGOING DISCUSSION with a different font as well to indicate the different nature). (I do not know if the query box implementation could be tweeked to give a Comments ’environment with one or two different features but in the meantime we could start putting comments in, flagging them up here in the forum.(<- and perhaps recordable in a separate list (Recent Comments) on the nLab, similar to Recently Revised perhaps)) I do think we are spending a lot of time talking here and perhaps can use existing technology to try out some ideas, so as to home in on what really is useful and what might be just superficial. The time is not wasted as such but might be used ’differently’. :-)
(Perhaps we could also try out approval tags on theorems using query boxes for the moment, as well.)
Couple of minor comments:
When I emphasise that stuff on the nForum should not be viewed as “permanent”, I don’t mean that it is in danger of being deleted! In terms of mere existence, the nForum is as likely to last as the nLab as far as I’m concerned. What I mean is that, by their very nature, discussions are difficult to follow afterwards and so if a conclusion is reached from a discussion then it should be summarised and put in the appropriate place on the nLab so that people coming afterwards don’t have to wade through reams of irrelevant stuff.
I regard nForum discussions as the “talk” pages of the nLab. The only piece of the jigsaw that I think is missing is the ability to find discussions by a particular page. That involves one bit of technology: making it easy to search by a particular nlab page name and only get links to nLab pages (not random discussions that happen to mention the words); and one bit of convention: that we use the page names when linking back (so that the search only has to be for one thing, not for all the potential redirects).
I really don’t like long discussions in query boxes. The query boxes should be only for when a casual reader of the page should know about the discussion - most aren’t of that form. Even then, I think that they get in the way somewhat. I would prefer that anything that hit the “scroll limit” of the query boxes should take place here and a link on the page should direct the reader to the discussion here if they so wanted.
With a bit of jiggery-pokery, it would be possible to put a “Discussion” link on each page that automatically linked to the discussions here on that page (once the search technology has been sorted out).
With a bit of jiggery-pokery, it would be possible to put a “Discussion” link on each page that automatically linked to the discussions here on that page (once the search technology has been sorted out).
This has been a long-held wish of mine. I think it would be great.
I really don’t like long discussions in query boxes.
Me neither!
There are in fact still a few long-forgotten discussion boxes sitting on various nLab pages that serve no purpose except making the page hard to read.
For either the discussion in the box achieved a worthwhile point, then that should be extracted into genuine entry text, or else the discusion resolved itself, in which case it ought to be removed anyway.
I get the point, and agree, so here is another tentative solution.. again to see if the idea works before getting into implementation details:
Use a query box which says (and this has been done before) ’ Comments: Discussion of …. can be found on the n-Forum at the link. …’.
The box will be small (and it need not be a query box and could be just part of the text) and the link would allow people to see the discussion at the n-forum. As I said there are pages where this is already done but we might do it more often and more consistently.
There are in fact still a few long-forgotten discussion boxes sitting on various nLab pages that serve no purpose except making the page hard to read.
I know some of my own and some of them I do not consider purposeless. They wait for another period in which I will be considering the same entries, probably in few months or so, and they will remind me or other discutant to reolsve the issue with further readings, thinking and so on. Why all this pressure here to hurry with finishing various things ? Actually the queries which I used much more in early days of nlab (i.e. before nForum times) were the advantage of nlab as opposed to record general knowledge on math in more quiet environment like wikipedia (Edit: only the second best advantage, the LaTeX is in nlab so much better). Have I not benefitted from them, I would probably not be very active in nlab at least until the days of nForum.
Toby said: But I would like that place to be the page itself, not a separate talk page, at least while the discussion is active.
This looks to me (Zoran) more or less like query boxes.
Tim: you ask us to temporarily do the computer's job. I am talking about the user perspective. The one who is concerned about finding discussion among hundreds from long period of time is the end user, not the creator of few discussions between today and few weeks from now when some of these issues will be likely partly implemented.
My point is merely that if we use comments such as ’ This is discussed in the n-forum [link] entries 58 - 67.’ with the link taking one at least to the page and better to the entry number. Once we have seen how that works out then we can gradually go and copy the discussions to new pages or whatever, editing them to increase clarity. If after a time the discussion is not needed in the lab then there would be very little editing to do to remove the link. This way the n-forum acts as the separate page for a while until the status is stabilised on the Lab page and the type of the comment is decided. It is a short term solution that may give us experience in how the long term implementation is best done.
With the ideas here, I think that it’s important to realise that there isn’t going to be a “one size fits all” solution. Some pages are very definitely “work pages” and on them, query boxes will almost be the norm: little notes left to oneself to remind oneself what one was thinking. But other pages are more exposition in purpose and on those I think that query boxes are detrimental. I’m thinking of things like: “I don’t know if this works constructively or not …”. Whilst people might be interested in such, it’s not something that is essential for the reader of the page to know. So for some pages, query boxes are best, for others, a link to an nForum discussion. Similarly, some discussions should be preserved and these should have their own nLab pages, but others were only necessary to resolve some issue and once it’s resolved, it’s done and isn’t needed again.
The main missing step is the easy search of and linking to the nForum. I’ll have a think about that.
Another thing to consider is whether or not we are considering a new type of discussion on pages. Should there be a “Talk” category on the nForum where these discussions go? They’re a little longer than what is normally logged in “latest changes”, but more focussed (and particularly focussed on a particular page) than the more general discussions that we sometimes have.
It would also be nice if Markdown had a comment syntax. We could probably rig something up with CSS visible=false, but it’d be better if the comment text just didn’t show up at all in the generated HTML.
I like the idea of a focussed page for discussion on the content etc of a given entry.
I’m not convinced about the comment syntax. The problem I have with it is that only if you put it there will you remember to look for it. Having it in the source isn’t exactly accessible (in its non-technical sense). If you want to leave yourself private messages, start a page “Mike Shulman: Thinks I Was Thinking” and make notes as you edit pages.
That said, certainly something with the CSS would be possible (just think of the drop-down lists). I had a couple of goes at seeing if I could get something in to the source that the sanitiser would strip out, but either it escaped it (meaning that it showed up) or it got caught by the spam filter.
Certainly a comment syntax could be abused. Where would you suggest that we make notes intended to be read by anyone else editing the page, but which may be irrelevant or confusing to someone merely reading it? The Forum discussion / Talk page is one possibility, but there is no guarantee that someone editing the page would think to look there for such a note.
I don’t think that I would particularly look in the source, either. I think that a discrete “Note to people editing this page:” at the bottom, followed by the comment or a link to a discussion here, wouldn’t be too disruptive and would be much more obvious - also, being at the bottom of the rendered page, it’s just where you’re going to see it before hitting “Edit”. Nowhere else is as likely to get noticed, as far as I can see.
I agree with Zoran about the usefulness of query boxes that have been lying around for a while. Some really are useless (because the issue has been resolved), but some are still useful.
I would very much not like to replace query boxes with a ‘Discussion’ link that searches the Forum (handy as that link would be). For some pages, this would work fine, but for others it would give me a list of several discussions, most of which are no longer current. I don’t want to have to read through them all to know what discussions are still relevant to the page as it is now. Instead, I would like to see a query box (or something like it) on the page telling me a human judgement of what discussion is still needed.
I don’t mind if long query boxes get replaced by links to the Forum as suggested by Tim in #8.
For some pages, this would work fine, but for others it would give me a list of several discussions, most of which are no longer current.
Which is part of my point about “one size fits all” solutions. Sometimes, query boxes are Just Right. Sometimes, a link to a specific discussion is the right choice. But also a large page may have many discussions and it’s useful to have a single link that finds all of them. It’s up to the author/editor at the time to decide which is most appropriate, but it’s up to us here and now to decide what the “standard” options should be. Indeed, perhaps the result of this discussion should merely be that:
It sounds good to me. I agree with some fine points in 16 as well; one sometimes has some unchecked material, remarks to reconsider etc. and wants to keep them in comment sense until later work and does not want really get now in query or some other discussion.
I don’t think that I would particularly look in the source, either [for a comment].
But if the comment were right there in the source next to the bit of text you planned on editing, you would see it without having to look for it anywhere.
I agree with the result suggested by Andrew #19. Except that I also agree with Mike that making an invisible comment (producing nothing in the rendered HTML, or producing an HTML comment if that’s easier) is also one of the sizes that might occasionally fit.
As for what we might type in the Forum that would be searched for when we hit the ‘Discussions’ link on the lab page, it would be handy if any redirect would work. For one thing, sometimes we change a page’s name, so always using the canonical name (at the time) will not always work. (Besides that, sometimes people misspell the name, and rather than tell them about it and get them to fix it, I usually just create a redirect for the misspelling.)
Any redirect or any pipelink as well, of course.
1 to 23 of 23