I don't know when this happened, but the rate at which questions are being asked seems to have increased dramatically relative to the last time I frequently browsed math.SE. It is getting hard to effectively filter through questions I don't want to look at to find questions I do want to look at. One thing that would help is the ability to hide questions I don't want to look at.
$\begingroup$
$\endgroup$
12
asked Feb 22, 2013 at 21:32
2 Answers
$\begingroup$
$\endgroup$
7
Added to confirm that Qiaochu isn't imagining things: the number of questions and answers per day. Questions in blue, answers in red. Time axis labeled by YY-MM.
That's right: over 300 questions and 500 answers per day...
Two reasons why this feature request is not going to be implemented.
Reason 1, political. This idea never got much traction on meta.SO, presumably because with their level of traffic, hiding individual questions is a Sisyphean task. Questions arrive faster than one can evaluate and hide them. And if a feature does not suit SO, it does not get implemented on StackExchange.
Reason 2, technical. Lists of questions are cached by the SE servers. All customizations of these lists (highlighting favorite tags, greying out or hiding ignored tags) are done by a client-side script after the page is loaded. This script (called full.js if you want to take a look) parses the content of the right sidebar and extracts the user's tag preferences from there. Amusingly, even 1 bit of information (whether to hide ignored tags) is served in the html as a hidden div just above the words "tag subscriptions" in the sidebar:
<div class="dno">
<input type="checkbox" id="hideIgnored" title="hide ignored tags" checked="checked"><label for="hideIgnored"> hide ignored tags</label>
</div>
This is where the script gets information on what to do with ignored tags. It then matches the tags in questions (which are CSS classes such as t-homework) and adds new classes such as tagged-ignored.
The upshot is that the list of questions is customized on the client side based on the information embedded in the markup of the page. Hiding individual questions would require sending a massive amount of data with the page, since there would be many more hidden questions than ignored tags. This is not practical.
A slightly more practical solution would be to keep the list of hidden questions locally in the browser. This is what the user-contributed script to which I linked in a comment attempts to do. I still haven't checked whether it actually works. And of course, this solution can work only if the user accesses the site from the same computer and browser.
Update. Julian and Asaf brought up the issue of how SE handles the questions that have both favorite and ignored tags. Assuming that "hide ignored" mode is enabled, the following happens: the questions gets highlighted and hidden from the page. In effect, "ignore beats favorite", as Julian said. This behavior does not fit the practice of scientific exchange. For example, a mathematician working in area A (but not in B) usually wants to know about applications of A to B and vice versa.
Unfortunately, the meta.SO proposal Provide “Keep Interesting Tags” option when “Hide Ignored Tags” enabled has not gained much support in over three years of existence. One answer is a link to a jQuery script, which I presume solves the issue for users who have GreaseMonkey or TamperMonkey in their browser.
Since I do not use all-purpose user-script extensions like TamperMonkey, I wrote a tiny single-purpose Chrome extension: it prevents questions with favorite tags from being hidden. If anyone wants to try it, download, unzip and install as unpacked extension.
The extension has no scripts. All it does is add a single CSS rule.
.tagged-interesting {
display: block !important;
}
This overrides the rule display: none; for .tagged-ignored-hidden Note that this works only on Questions and Unanswered tabs, not on the front page. That is, it affects the pages that show questions with two-line summaries, not the pages that show titles only. The titles-only lists are composed dynamically by a script that actually omits ignored questions (rather than hiding them with CSS).
But question lists with two-line summaries are more useful anyway: no need to open a question in which the first word is Prove, Show or Solve.
-
$\begingroup$ By the way, this implementation of tags-as-CSS-classes is a roadblock for internationalization of SE sites. Since class names can contain only a very restricted set of characters, it is impossible to use international characters in tags. This is why some Area51 proposals are stuck with this message. $\endgroup$– user53153Feb 23, 2013 at 16:56
-
$\begingroup$ Interesting. Do you know this from having worked at SE, or was it inferred from monkeying around with scripts? $\endgroup$ Feb 24, 2013 at 1:02
-
2$\begingroup$ @MathGems I don't work at SE, I just press F12 in Chrome a lot. $\endgroup$– user53153Feb 24, 2013 at 1:10
-
$\begingroup$ It is not true that features need to be useful on SO to be implemented. Quite a few sites have site-specific modifications, e.g MathJax, Balsamiq on UX, chemical formula support on Bio and Chem and more. The "Citation needed" post notice e.g. was requested by Skeptics and implemented, though it isn't of much use to most other sites. $\endgroup$– user9733Feb 24, 2013 at 12:40
-
$\begingroup$ @MadScientist That's true, site-specific addons exist. My post was in the context of the proposed interface change, which (I imagine) would be done network-wide if at all. And such network-wide changes would not happen without SO being on board. $\endgroup$– user53153Feb 24, 2013 at 16:34
-
$\begingroup$ I suppose that the sharp decrease is Christmas and holidays? :-) $\endgroup$– Asaf Karagila ModFeb 26, 2013 at 23:04
-
$\begingroup$ I don't want to bump this post by changing the Chrome extension link, but I'll point out that a more useful extension is available elsewhere. $\endgroup$– user53153Mar 2, 2013 at 22:28
$\begingroup$
$\endgroup$
14
Partial solution: I bit the bullet and started hiding a few tags I was pretty sure I generally didn't want to look at. This lets me hide some recent questions by tagging them with those tags if relevant.
answered Feb 23, 2013 at 2:03
-
7$\begingroup$ Retagging questions to hide them (even if it does fit) is really a bad abuse of the tag system. I am all against it. $\endgroup$– Asaf Karagila ModFeb 23, 2013 at 2:09
-
2
-
5$\begingroup$ Because it feels to me that it promotes sort of using the tags for your own personal gain, rather in promotion of the general [well-]order of questions on the site. It feels a bit like running for moderator only to be able to edit your comments past the five minutes window. $\endgroup$– Asaf Karagila ModFeb 23, 2013 at 2:11
-
4$\begingroup$ But tagging questions with relevant tags does promote the general well-order of the site. This particular way of doing it happens to also save me some time when browsing questions (and anyone else who's hiding the same tags as me). If I wanted to abuse the tag system I would invent a tag solely for the purpose of hiding questions (which, to be clear, I have no intention of doing). $\endgroup$ Feb 23, 2013 at 2:12
-
2$\begingroup$ I still have a feeling that this is sort of a "everybody wins" gray area exploitation, which inevitably leads to some sort of a "everybody loses" outcome. I hope I'm wrong though. $\endgroup$– Asaf Karagila ModFeb 23, 2013 at 2:15
-
$\begingroup$ Firstly, I just went to verify a feature of the hiding process: even if a question is tagged under your favorite tags and under an ignored one, it will be hidden. But more importantly, I feel that this sort of attitude (if becomes more common) can lead to some bias in retagging, bias which may ultimately cause harm to the tag system. (And I agree, a flat out abuse would be to create [qchu-hates-this] tag, but that would obviously get removed pretty fast...) $\endgroup$– Asaf Karagila ModFeb 23, 2013 at 2:17
-
7$\begingroup$ @AsafKaragila So far I don't see Qiaochu suddenly retagging a bunch of questions to hide them. Let's keep an eye on this activity list. :) // Anyway, I don't see hiding-motivated retagging as inherently wrong. A hypothetical scenario: someone hates PDE and hides the
(pde)tag. When they see a PDE problem, they say 'Yuck! Why do I see this? Aha, the tags are wrong." -- and the problem falls into one of my favorite tags. Win-win. $\endgroup$– user53153Feb 23, 2013 at 2:44 -
$\begingroup$ @5pm: Call me a pessimistic fatalist, I'm not going to fight too much about this issue. I'm just pointing out that it does have a pretty big loophole. Perhaps by brining it up I caused more people to observe this and I will be the result of the future abuse. Ah, good ol' determinacy, contradicting the very idea of choice. I hope that we all do agree that a loophole does exist in this method, much like several other by-design-holes that I am aware of in the system (which I am sometimes using, and sometimes not; but never revealed). $\endgroup$– Asaf Karagila ModFeb 23, 2013 at 2:49
-
$\begingroup$ "I hope that we all do agree that a loophole does exist in this method" -- Sorry, @Asaf, I've read your comments multiple times but I still don't understand what this hypothetical loophole that you keep alluding to actually is or what harmful effect it would have. Can you be a little more specific? Are you saying people would start retagging questions they don't like with the wrong tags just to hide them? $\endgroup$– user856Feb 23, 2013 at 3:20
-
$\begingroup$ (Meanwhile, 5pm has provided a concrete benefit that this sort of retagging would have, so as far as I'm concerned the score is 1-0 in favour.) $\endgroup$– user856Feb 23, 2013 at 3:22
-
2$\begingroup$ @ℝⁿ. I guess I'm probably more paranoid than others. I'm not saying people will tag things wrong, but rather they start to develop bias towards particular tagging habits, and that bias is not always going to serve the best interest of the community; and as a pessimist I feel that eventually this is not going to be for the better. $\endgroup$– Asaf Karagila ModFeb 23, 2013 at 13:45
-
5$\begingroup$ @5pm Out of curiousity you've only been a member for 2 months how come you know so much about how things go? $\endgroup$– user38268Feb 23, 2013 at 23:41
-
1$\begingroup$ @BenjaLim I've got amazing powers of observation. $\endgroup$– user53153Feb 24, 2013 at 0:31
-
$\begingroup$ @5pm Some phenomena can only be explained by reincarnation. $\endgroup$ Feb 27, 2013 at 8:43
...even if they also match one of my interesting tags." This is a quote from meta.SO, actually. // A simple client side script could fix this too. But having the second checkbox would also alert the users to the fact that they may be ignoring some of their favorite-tagged questions. At present the users are not really informed about the effect that their preferences have. $\endgroup$