Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Site Tag Cloud

2-category 2-category-theory abelian-categories adjoint algebra algebraic algebraic-geometry algebraic-topology analysis analytic-geometry arithmetic arithmetic-geometry book bundles calculus categorical categories category category-theory chern-weil-theory cohesion cohesive-homotopy-type-theory cohomology colimits combinatorics complex complex-geometry computable-mathematics computer-science constructive cosmology definitions deformation-theory descent diagrams differential differential-cohomology differential-equations differential-geometry digraphs duality elliptic-cohomology enriched fibration foundation foundations functional-analysis functor gauge-theory gebra geometric-quantization geometry graph graphs gravity grothendieck group group-theory harmonic-analysis higher higher-algebra higher-category-theory higher-differential-geometry higher-geometry higher-lie-theory higher-topos-theory homological homological-algebra homotopy homotopy-theory homotopy-type-theory index-theory integration integration-theory k-theory lie-theory limits linear linear-algebra locale localization logic mathematics measure-theory modal modal-logic model model-category-theory monad monads monoidal monoidal-category-theory morphism motives motivic-cohomology nforum nlab noncommutative noncommutative-geometry number-theory of operads operator operator-algebra order-theory pages pasting philosophy physics pro-object probability probability-theory quantization quantum quantum-field quantum-field-theory quantum-mechanics quantum-physics quantum-theory question representation representation-theory riemannian-geometry scheme schemes set set-theory sheaf simplicial space spin-geometry stable-homotopy-theory stack string string-theory superalgebra supergeometry svg symplectic-geometry synthetic-differential-geometry terminology theory topology topos topos-theory tqft type type-theory universal variational-calculus

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to nForum
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorMike Shulman
    • CommentTimeJan 1st 2010

    Cross-web links in instiki currently look kind of ugly unless you make them into pipelinks. I think they look much nicer on the n-Forum. For example, on the Forum, [[michaelshulman:HomePage]] comes out as HomePage (michaelshulman) instead of michaelshulman:HomePage as it does in Instiki. If, as seems likely, cross-web links are going to become more common, I'd like to ask Jacques whether Instiki's default display of cross-web links could be prettified. Opinions?

    • CommentRowNumber2.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 4th 2010

    Jacques has just changed this. It seems that the lab name is now encoded as the title attribute of the link and there's some CSS magic that puts a box around links to other webs. Take a look at the last sentence at HomePage (doriath) to see how it looks.

    Changing the CSS would be simple, putting the web name back into the running text would require a deeper fix.

    Jacques did say (commenting on this change) "Feedback welcome".

    • CommentRowNumber3.
    • CommentAuthorTobyBartels
    • CommentTimeJan 4th 2010

    Yeah, those boxes are ugly (which, as you say, is easy to change with CSS). The overall systems seems sound to me.

    • CommentRowNumber4.
    • CommentAuthorTobyBartels
    • CommentTimeJan 4th 2010

    The boxes look particularly wild at http://ncatlab.org/web_list.

    • CommentRowNumber5.
    • CommentAuthorUrs
    • CommentTimeJan 4th 2010
    • (edited Jan 4th 2010)

    These boxes on cross-web links currently mess up the look of essentially all the pages on my personal web, where I have plenty of cross-web links. Hopefully there is a way to make them disappear.

    My impression is the following might be important:

    there are really two different ways that hyperlinks on a wiki are used

    1) as explicit links that ask the reader to follow them. Something that in the text would be accompanied by a statement like "For more details see ...".

    2) As just indications that further material on a given keyword is available, if desired, but being an option that the reader generally may ignore.

    I am personally using extensively hyperlinks in the sense of point 2 and I think most of the links in the nLab are of this kind.

    Now, type 2 links should not be highlighted much. Just barely, so that one can tell whether or not a keyword has a hyperlink to it or not. An underline as we used to have is okay. A slight shading of the word, that becomes brighter as one mouses over it, as on the Cafe, might also be an option.

    On the other hand, type 1 links want to be clearly highlighted. Maybe boxed, as we have now. As these are really placeholders for text to be found elsewhere, they need not look like ordinary text.

    So if we could distinguish between two kinds of hperlinks in the source code, we should adapt their highlighting accordingly. Since we can't I think we should have highlighting that reflects the majority of usage of links, and for us this is type 2.

    • CommentRowNumber6.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 4th 2010

    Yes, that is ugly. To fix that instance we'll need to change the main instiki.css file rather than doing it on a web-by-web basis. Still simple, though. Has anyone a good suggestion on how to differentiate between in-web and inter-web links?

    • CommentRowNumber7.
    • CommentAuthorUrs
    • CommentTimeJan 4th 2010
    This comment is invalid XHTML+MathML+SVG; displaying source. <div> <blockquote> Has anyone a good suggestion on how to differentiate between in-web and inter-web links? </blockquote> <p>Do we want a distinction? I thought it was nice that there is none.</p> </div>
    • CommentRowNumber8.
    • CommentAuthorMike Shulman
    • CommentTimeJan 4th 2010

    I think there should be a distinction. People should be made aware of when they are moving from (say) the main nLab to a personal web, or vice versa. What about an icon before or after the link, similar to mediawiki's "external link" icon?

    • CommentRowNumber9.
    • CommentAuthorMike Shulman
    • CommentTimeJan 4th 2010

    I am also against trying to distinguish in display between Urs' "2 uses of hyperlinks," partly because that would necessitate modifying the syntax for creating at least one of them, and I think simple and easy-to-remember link syntax is good.

    • CommentRowNumber10.
    • CommentAuthorTobyBartels
    • CommentTimeJan 5th 2010

    @ #8

    Would a difference in colour be enough? (Technically something involving both saturation and hue, for good contrast normally but also some contrast for colour-blind readers.)

    • CommentRowNumber11.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 5th 2010

    Jacques has put in an update on the stylesheet that removes the boxes from (most) of the links on the web list.

    I think that distinguishing between Urs' two categories (sorry) is difficult from a technical side. It's perfectly possible if authors put in some flag that says "type 1" or "type 2" but are people really going to do that?

    I also think that there should be some distinction between within-web and other-wed links and that it should be obvious that there's a difference to someone encountering the nLab for the first time.

    • CommentRowNumber12.
    • CommentAuthorUrs
    • CommentTimeJan 5th 2010

    Thanks, Andrew.

    By the way, I didn't propose that we should distinguish in the source code between the two types of links. I just pointed out that these two different usages of links exist, and that this should be kept in mind when designing their appearance.

    The boxes were thought up by somebody who had links of tye 1 in mind. For us, who use mostly links of type 2, that doesn't work. That was just my point.

    • CommentRowNumber13.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 5th 2010

    Ah, okay. I understand now.

    Well, so the basic problem is still extant: we don't like the current style. So we need a concrete suggestion. If it involves showing the web name, that'll need code hacks, but if it's just style then we can do that with CSS. (Just to be clear: the web name is encoded in the 'title' attribute so shows up if you hover over the link - at least, with firefox it does.)

    What would people prefer?

    • CommentRowNumber14.
    • CommentAuthorUrs
    • CommentTimeJan 5th 2010
    • (edited Jan 5th 2010)
    This comment is invalid XHTML+MathML+SVG; displaying source. <div> <blockquote> If it involves showing the web name, that'll need code hacks, but if it's just style then we can do that with CSS. </blockquote> <p>Let's not make the web name show. That doesn't make sense for links of type 2 and in fact would spoil type 2 usage.</p> <p>Here on the forum, we exclusively use links in the sense of type 1. Accordingly, here it does makes good sense to include the web name. But on the wiki, it would break the type 2-practice.</p> <p>I would prefer if we simply went back to the original behaviour.</p> </div>
    • CommentRowNumber15.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 5th 2010

    The original behaviour did show the web name (see Mike's first comment). The new behaviour hides the web name slightly so you only see it when you hover over the link. However, I think that there should still be some slight visual clue that links to a different web are different to the links to the current web. The boxes are meant to indicate that, but we don't seem to like the boxes so we need some other indication. The easiest methods are with colours, say by shading the background or darkening the colour, or we can put a border around it or under it; it's also possible to put symbols before or after, as in Mike's suggestion. Pretty much anything is possible with CSS so just suggest something and I'll see if it's possible.

    • CommentRowNumber16.
    • CommentAuthorUrs
    • CommentTimeJan 5th 2010
    • (edited Jan 5th 2010)
    This comment is invalid XHTML+MathML+SVG; displaying source. <div> <blockquote> The original behaviour *did* show the web name (see Mike's first comment). </blockquote> <p>Oh, sorry, I mixed up piped and un-piped links (I gather that's how the are called).</p> <p>Ah, so that wasn't even clear to me: even un-piped links to another web now no longer spell out the full web name in the text. Ah, that's good.</p> <p>Concerning design: since the main distinction between different webs is the <strong>color</strong>, maybe we could have a cross-web link take the color of the web that it links to. That would make it very intuitively clear for a user, say, in the green nLab zone, that a link will take him to a red personal zone.</p> <p>(In that case we might want to harmonize personal web color schemes.)</p> </div>
    • CommentRowNumber17.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 5th 2010

    I wondered whether or not someone would suggest that!

    Not really very practical, I'm afraid. That would require adding another database look-up to find out the colour for each link.

    Also, I'm not convinced that's useful. There are (currently) only a small number of available colours so there's bound to be some repetition. And a page could potentially have a riot of colours with no obvious explanation (to the casual visitor) as to what the different colours mean.

    • CommentRowNumber18.
    • CommentAuthorUrs
    • CommentTimeJan 5th 2010

    Right. But what if we simply adopted nLab=green and all personal webs=red? Just so the user knows whether he/she's in the public or in a personal area?

    • CommentRowNumber19.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 5th 2010

    Yes, that's possible from the nLab side. It wouldn't work from the personal web side (since there one has three categories: this web, the nLab, and another personal web) but then people with personal webs can decide on their own colour settings.

    I've implemented this on doriath (see HomePage (doriath)) just to check that it works and to see what it looks like (I'd rather just do one change on the nLab itself rather than keep changing it every day). Feel free to add more examples of inter-web links there to see what it looks like more extensively.

    • CommentRowNumber20.
    • CommentAuthorUrs
    • CommentTimeJan 5th 2010

    Andrew, would there be a chance that you could make a quick implementation of some such alternative link formatting to my personal web. For instance the one you just implemented on doriath.

    I regularly point people to various parts of my personal web, for various purposes, and currently my whole web looks sort of broken, because each page presents a myriad of boxes to the viewer. I need to get rid of these boxes as soon as possible.

    If you give me a hint for what I need to do to change it myself, I'll try it.

    • CommentRowNumber21.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 5th 2010

    How's that?

    Technically, it's "saddle brown". I chose it as a bit darker than orange.

    • CommentRowNumber22.
    • CommentAuthorUrs
    • CommentTimeJan 5th 2010

    Thanks, Andrew!!

    • CommentRowNumber23.
    • CommentAuthorMike Shulman
    • CommentTimeJan 5th 2010

    I disagree that showing the web name breaks type 2 links. Showing them in the way they used to be shown for non-piped links did break them, which is one reason I wanted to change it. But I don't think that showing them in parentheses afterwards is a problem. Parentheses are ignorable.

    I also like personal webs being able to choose their own colors. (-:

    • CommentRowNumber24.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 6th 2010

    Putting the web names in parentheses will involve a (minor) hack in the code. It's perfectly possible, but is slightly more major than just changing the style.

    So as a first solution, if owners of personal webs want to suggest their own colours for 'intra-web' wikilinks, I'll change all the CSS to reflect those.

    I agree that if the web names are shown, then they should be in parentheses for exactly the reason that Mike says. I think that they should be shown as it makes it obvious to a casual visitor what is going on. Obviously, anything that we know about is going to work for us but we need to keep an eye on how it is perceived from the outside.

    • CommentRowNumber25.
    • CommentAuthorUrs
    • CommentTimeJan 6th 2010

    would it be possible to have the web name in paranthesis and sort of lik a subscript, i.e. in smaller font.

    I am just afraid that on a page where every 5th word is hyperlinked, and that to another web, dispaying all these web-names will look very odd.

    Don't get me wrong, I certainly don't intend to trick anyone into believeing a link goes somewhere it doesn't go. I am just afraid that with links used abundantly, marking each link by extra text will result in weird typesetting.

    What if we make the mouse-over tag more prominent? So that everyone about to click on the link can't fail to be alerted as to where the link points?

    • CommentRowNumber26.
    • CommentAuthorMike Shulman
    • CommentTimeJan 6th 2010

    By "choose their own colors" I meant just choose their own colors for the personal web itself. (there was a suggestion made that "all personal webs become red").

    • CommentRowNumber27.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 6th 2010

    @Mike: The phrase "all personal webs become red" referred only to wikilinks to personal webs (at least, that's how I interpreted it).

    @Urs: I'm wary about smaller fonts. I suspect that these would actually draw attention to the webname rather than the opposite. For paragraphs with lots of links, it's worth remembering that the [[web:page name|some random text]] syntax will always override any defaults.

    • CommentRowNumber28.
    • CommentAuthorMike Shulman
    • CommentTimeJan 6th 2010

    Oh, ok.

    Whether pipelinks override whatever indication we choose could also be regarded as negotiable. The current "box" syntax, for instance, applies to pipelinks as well as un-piped ones. One might argue that the casual visitor should be warned when they are moving from one web to another, whether or not the link is piped.

    • CommentRowNumber29.
    • CommentAuthorTobyBartels
    • CommentTimeJan 6th 2010

    @ Andrew #19

    I don't think that there are any links from one personal web to another. Of course, someday there may be, but they'll probably still be rare.

    • CommentRowNumber30.
    • CommentAuthorTobyBartels
    • CommentTimeJan 6th 2010

    With the interweb links on Doriath, I see no difference between links to pages that I've visited and links to pages that I haven't visited. It may be very Web 1.0 of me, but I find that difference useful.

    • CommentRowNumber31.
    • CommentAuthorEric
    • CommentTimeJan 7th 2010

    For what its worth, I don't think the old way was broken and don't understand why we needed to fix it. I like the way the nForum handles links, but that doesn't mean we needed to change the nLab.

    My vote: Change it back to the way it was originally on the nLab.

    • CommentRowNumber32.
    • CommentAuthorMike Shulman
    • CommentTimeJan 7th 2010

    @Eric, with our new terminology, I would say that the problem with the old way was that type-2 links are impossible to make between webs unless they are pipelinks. That is kind of a pain if you are doing a lot of linking between webs.

    • CommentRowNumber33.
    • CommentAuthorUrs
    • CommentTimeJan 7th 2010

    I agree with Eric that the previous state of affairs was good as far as output went, and with Mike that it was a pain that one had to pipelink to get that.

    So why not have simply the best of both worlds: have cross-web unpiped links display as just ordinary links.

    I would prefer that. Just plain links everywhere.

    • CommentRowNumber34.
    • CommentAuthorEric
    • CommentTimeJan 7th 2010

    I second that

    • CommentRowNumber35.
    • CommentAuthorMike Shulman
    • CommentTimeJan 7th 2010

    Okay, so we seem to have a fundamental disagreement here. Andrew and I think that the web should be visible in a cross-web link, so as to make it clear to the casual visitor what's going on. Urs and Eric think it should not be visible -- I gather mainly because it's ugly, especially for type 2 links? How would you two feel about a small icon appearing next to all cross-web links?

    • CommentRowNumber36.
    • CommentAuthorUrs
    • CommentTimeJan 7th 2010

    That's a correct summary as far as I am concerned.

    I happen to have pages that use cross web links abundantly, so naturally I am getting nervous with designs that would make these pages look very ugly (even more ugly than they are anyway...)

    A small icon might be an option. What else could we try? How about this: currently links are identified by color and underlining. What if we use both separately? Just colored for one type of link, colored and underlined for another?

    • CommentRowNumber37.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 7th 2010
    Right, so it seems that there's not enough support for the "web name must be visible" view to warrant hacking the code. That's fine. (Of course, we can always revisit this later.)

    So that brings us to CSS which is easy to do and easy to try out. One thing not to lose is that this can be done on a web-by-web basis. I would say that it is more important that it be obvious to a reader that they are going "off web" on the nLab itself than on a personal web. So I can see the argument for, say, Urs to have on his personal web the difference just down to colour but on the main nLab having an icon as well.

    I also agree with Toby's point about visited and unvisited links, but again this is CSS.

    So we need a scheme. Someone please propose a scheme and I'll implement it, first on doriath to see what people think. What we need are:

    1. Unvisited links in same web
    2. Visited links in same web
    3. Unvisited links to other web
    4. Visited links to other web

    CSS can do lots: colours (fore and back), add "tags" and "icons" before or after elements, add borders, change font styles (size, weight, etc). But it's not a programming language so it can do "if"s and "choose according to"s.

    Also, if personal web owners know what they want then just say what you want and we'll try to do it.

    Don't worry about getting the CSS syntax right - just come up with suggestions.
    • CommentRowNumber38.
    • CommentAuthorEric
    • CommentTimeJan 8th 2010

    In my opinion, we should not be concerned about identifying cross links. Why should we care? A malicious person trying to fool someone? On the nLab? Really?

    There are some rare occasions where we want the cross link to be specific, but in those rare cases, why don't we just use a pipe link? There are enough lab elves around to keep the place clean without messing up Urs' personal web.

    In my opinion, aesthetics are important and I think all links within a given web should be uniform.

    Having said that, I do like the links the way they are here on the nForum, but the nForum is different.

    • CommentRowNumber39.
    • CommentAuthorTobyBartels
    • CommentTimeJan 8th 2010

    I would like it if all of the links on Urs's web look the same, but links to it from the main web were identifiable. (Which is how it was before, really, since Urs used pipe links on his web but not on the main web.)

    I suggest using the link colour scheme for regular links on Urs's web (which is different from the scheme on the main web) for cross-web links on the main web. Cross-web links on other webs should look normal (whatever that is on that web) by default.

    • CommentRowNumber40.
    • CommentAuthorEric
    • CommentTimeJan 8th 2010

    Or...

    Just put pipe links on the nLab that point to Urs' personal web. There aren't that many and I don't think it is worth messing up stuff.

    Are we going to do the same special case for my personal web? I like my links the way they were too.

    • CommentRowNumber41.
    • CommentAuthorTobyBartels
    • CommentTimeJan 8th 2010

    In my suggestion for #39, it is the main lab that is the special case. On personal webs by default, there would be no distinction made between links, although the owner could certainly request something. Only on the main web, I liked having the distinction (done by a prefix there, but colour would be fine).

    • CommentRowNumber42.
    • CommentAuthorEric
    • CommentTimeJan 8th 2010

    This is an interesting situation. How do we resolve it? A vote?

    I think I would dislike seeing multicolored links (or marking of any kind) on the nLab as much as I dislike seeing "s" outside brackets :)

    I think the default behavior for cross links should be no markings and no color differences. If we want to make a particular link clear that it is a cross link, we can opt to provide a pipe link.

    • CommentRowNumber43.
    • CommentAuthorMike Shulman
    • CommentTimeJan 8th 2010

    Eric, the question is not about "malicious" users, but simply about readers who are ignorant. They may not even realize there are separate webs. They should be notified in some way when they move from the main nlab, where lots of people scrutinize every word, to Joe Wacko's private web where he writes whatever the heck he feels like. (Okay, that's an exaggeration, but you get the picture.)

    I don't really like the idea of separate colors for different kinds of links. Changing font styles would also be ugly, I think. I propose keeping colors as they are (separate colors for visited/unvisited only), but adding small icons after cross-web links on the main lab. I don't think this should produce any extreme ugliness, since there are really quite few cross-web links on the main lab, and even fewer are of type 2. And letting owners of personal webs make their own choices about how links look there makes perfect sense.

    • CommentRowNumber44.
    • CommentAuthorMike Shulman
    • CommentTimeJan 8th 2010

    BTW, of course Urs is not Joe Wacko. (-: If anything, I'm thinking more of myself. I wouldn't want anyone to end up on my personal web thinking that it is part of the main nlab.

    • CommentRowNumber45.
    • CommentAuthorUrs
    • CommentTimeJan 8th 2010
    • (edited Jan 8th 2010)

    That's fine Mike, I did understand you! :-) In fact, I want people to be alerted when they come to my web just as much as anyone might want this. After all, much of the point of having my personal web is precisely to say to people: "hey, here is material I am thinking about, if you have any comment, please don't hesitate to tell me".

    And it's also a matter of claiming idea ownership, isn't it? If I put my idea/observation/theorem on the nLab, I sort of give it away to the public domain more than when I mention it on my personal web.

    So I am all in favor of identifying cross-web links. But it has to be such as not to break everything else.

    But we can treat the main web differently from the personal webs, right? So by all means, let's put icons or whatever to cross-web links FROM the main web, and worry about what to do about cross-web links TO the main web later.

    • CommentRowNumber46.
    • CommentAuthorUrs
    • CommentTimeJan 8th 2010
    This comment is invalid XHTML+MathML+SVG; displaying source. <div> <p>Toby wrote:</p> <blockquote> I would like it if all of the links on Urs's web look the same, but links to it from the main web were identifiable. (Which is how it was before, really, since Urs used pipe links on his web but not on the main web.) </blockquote> <p>Yes, exactly. That would be good.</p> </div>
    • CommentRowNumber47.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 17th 2010

    I've done the main nLab the same as doriath: links off the nLab are red. This is a bit of a stop-gap since we haven't sorted out the full colour scheme according to the four things in my list above. However, as it seems no-one likes the boxes at all then I thought it best to remove them altogether.

    I've done Eric's web as well since he asked for it elsewhere. Anyone else who wants their web done, just ask.

    • CommentRowNumber48.
    • CommentAuthorTobyBartels
    • CommentTimeJan 17th 2010

    I think that this is the current implementation of the four-colour scheme on the main web (and Doriath):

    1. Unvisited links in same web: Green
    2. Visited links in same web: Gray
    3. Unvisited links to other web: Red
    4. Visited links to other web: Red

    Yes?

    If so, then I suggest this scheme:

    1. Unvisited links in same web: Green
    2. Visited links in same web: Gray
    3. Unvisited links to other web: Orange
    4. Visited links to other web: Red

    (This changes #3, rather than #4, to keep with the standard that visited links are darker.)

    I don't know how to write that in CSS; I'm assuming that you do, Andrew, which is why I'm not looking it up. (^_^)

    • CommentRowNumber49.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 17th 2010

    There you go. It's more of a reddish-brown than orange, though. The orange colour that I found was far too bright so I dulled it down and that was my second attempt. If anyone wants to suggest an alternative then please do (but it does help if you specify particular RGB values rather than just "orange"! There are quite a few variants of "orange" to pick from!). I've done the same on doriath (as that's the test web) but haven't done anything on any others.

    • CommentRowNumber50.
    • CommentAuthorTobyBartels
    • CommentTimeJan 17th 2010
    • (edited Jan 17th 2010)

    I thought that you were just using the colours given in the Instiki colour options; that's why I used names.

    • CommentRowNumber51.
    • CommentAuthorTobyBartels
    • CommentTimeJan 17th 2010

    Now that I have the password (^_^), I'm inclined to change *all* of the personal webs (except the ones that already have treatments at the request of their owners) to the same system that Eric has. That is, I suggest that system as the default, with the option for something different by request as usual.

    • CommentRowNumber52.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 17th 2010

    I was tempted to do the same, but am out of time now. I think that's completely reasonable and since it's easy to fix, shouldn't cause anyone to get annoyed with us (especially since the change to boxes was foisted on us all by a higher - extremely benevolent - power!).

    (And I didn't think about looking at the RGBs for the instiki colours. Completely missed that!)

    • CommentRowNumber53.
    • CommentAuthorTobyBartels
    • CommentTimeJan 18th 2010

    It now looks strange to me that, while we have (on many webs) special colours or marks for links to other webs, links to completely different sites look just like links to the same web!

    So now I have a 3-colour (really 6-colour, because of visited vs unvisited links) system on my web; all 3 appear on HomePage (tobybartels).

    • CommentRowNumber54.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 19th 2010

    Good point! If we want to signal when a user is moving from one web to another, we should do the same for when they are going off-site. Colour suggestions for off-site links on the main nLab?

    (I couldn't easily distinguish between the colours for off-site and within-web links on your homepage.)

    • CommentRowNumber55.
    • CommentAuthorTobyBartels
    • CommentTimeJan 19th 2010

    I couldn't easily distinguish between the colours for off-site and within-web links on your homepage.

    That's because I coloured my web Purple. I changed it to Green to match the main web, so how does it look now?

    • CommentRowNumber56.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 19th 2010

    Yes, that's clear.

    • CommentRowNumber57.
    • CommentAuthorTobyBartels
    • CommentTimeJan 20th 2010

    OK, I've put those coloured external links on the main web now (also Doriath, for good measure). Complain (or change them) if they're ugly.

    • CommentRowNumber58.
    • CommentAuthorMike Shulman
    • CommentTimeJan 22nd 2010

    Looks nice to me. I want this for my personal web.

    • CommentRowNumber59.
    • CommentAuthorAndrew Stacey
    • CommentTimeJan 22nd 2010

    @Mike: same colour scheme as the main nLab, or do you want it more in keeping with the colour of your web at the moment? Toby's scheme is that there are three main colours: within-web links, other web links, and off-site links. Then "visited" links get a darker version of the main colour. So to be sure what you're asking, could you say what you want those three main colours to be?

    • CommentRowNumber60.
    • CommentAuthorMike Shulman
    • CommentTimeJan 23rd 2010

    Same as the main nLab would be fine. I think making this the default would make sense -- at least, something other than the ugly boxes ought to be the default.

    • CommentRowNumber61.
    • CommentAuthorUrs
    • CommentTimeJan 25th 2010

    currently when going with the mouse pointer over a cross-web-link, it gets colored in a way that looks almost like it is being blackened by a black bar.

    This is not so good, I think. We should to use much brighter colors here, if we want such an effect at all.

    is it possible to have it the way it is at the nCafe, that not the background of the link text changes color, but just the link text itself, when "mousing over" it?

    • CommentRowNumber62.
    • CommentAuthorTobyBartels
    • CommentTimeJan 25th 2010

    There have been other unhappy CSS changes recently. Here is another.

    We might want to go over the whole thing (links, that is) and set it up nicely.