This is essentially a follow-up to Don't Retitle This Node.
Today, someone came along and asked a question about printing line numbers for use with debugging. He titled his node program line&dagger. Before long, Old_Gray_Bear considered the node for retitling with the suggested title "How Do I Get the Line Number In My Error Message?"
Which do you think is better? Why?
I think the original is better. By far. Here's why...
If someone else comes to Perlmonks with the same question, sees the search box on the top left, and puts in the words "program line", he's going to go directly to that node and find his answer. As it seems that search term is a good one for the question, that's a desirable outcome.
If we retitle the node to the suggested replacement, however, the search term "program line" will, as far as I can tell, pull up a big list of nodes, of which none deals with this issue. Not only that, but relevant searches that will find this node, like "error line" or "line number" already turn up nodes in which similar questions are asked and all the right answers are given.
What I'm getting at here is that most nodes shouldn't be retitled... even if you think their titles are too short or too cryptic. Remember, the poster seemed to think their chosen title was relevant, so its likely that someone else will eventually visit with the same problem and the same words in his mind. And when he searches, he'll find an answer.
Keep in mind that monks who have reached the level where they can consider a node tend to have a bit more experience than many of the people who come asking questions. And so, we are likely to think about those problems differently and draw different word associations when describing them. It is precisely because we don't look at those problems the same way that we aren't the best choice to pick a title. Afterall, we're not the ones searching for the answers.
So, please... think twice before considering a node for retitling. Don't do it for aesthetics, to be explicit, or because you think the wording isn't quite right. Don't do it if the title is merely bad. Even misspellings should probably be left as is because if someone misspelled or typo'd it in his title today, then someone else will probably misspell or typo it in the search box tomorrow.
The only nodes that should be retitled are ones that are downright unarguably terrible. Completely generic titles like "i have a problem" or "HELP!!!" are fair game. Profanity should probably be eliminated. Getting rid of excess punctuation might be prudent. There are probably other good reasons but, off the top of my head, those are about the only ones I can think of. If you are considering for another reason, take a moment to pause and consider why first.
† Update: This node has, unfortunately, since been retitled. I'm leaving the original link as a demonstration of yet another reason why frivolous node retitling is bad: it can leave a mess. It also demonstrates that the node doesn't even exist in the results returned by searching on that term now. For those really interested in the original node, it's node_id is 505311.
-sauoq
"My two cents aren't worth a dime.";
Re: Consider this: What makes a good node title?
by jdporter (Paladin) on Nov 03, 2005 at 23:11 UTC
|
But what if someone comes in and types "error line" in the search box? Or "line number"? Don't they deserve to find this node as well?
IMHO, the only way OGB's suggested title could be improved would be to insert "Program" in front of "Line". Generally, longer titles are better, at least if the words are high in information content.
someone else will probably misspell or typo it tomorrow
That's an extremely bad reason for retaining a misspelling.
It would foster the growth of subcommunities that share misconceptions about certain words. E.g. we'd have the "occurance" people sharing information, apart from the mainstream "occurrence" crowd, because they wouldn't find each other's nodes. (I'm serious about this.)
I don't think it's a problem to fix misspellings in titles, because generally the same misspelling occurs in the body of the node — and since janitors don't fix such things in node contents, a Super Search on the misspelling will find the node.
IMHO, one of the main reasons for retitling nodes is as a pedagogical instrument: the janitors are showing the OP, and others, what constitutes a good title for a given question.
We're building the house of the future together.
| [reply] |
|
Don't they deserve to find this node as well?
Are you asking me if they deserve to be inundated with duplicate information? Or if they deserve to be treated to this particular cookie of a node?
What they deserve is to find their answer (no matter which words come to mind) and more people will if we tamper with titles less.
E.g. we'd have the "occurance" people sharing information, apart from the mainstream "occurrence" crowd, because they wouldn't find each other's nodes.
Sorry, but that's a ridiculously bizarre notion. I know how to spell "occurrence" correctly but I wouldn't hesitate for one second to reply to a node with "occurrance" in the title. I don't think anyone else here would either. (At least, not most.) But, if someone believes it's a good search term for their problem, they should have a reasonable chance of finding a relevant node by searching on it regardless of whether they misspell it.
the janitors are showing the OP, and others, what constitutes a good title for a given question.
Nobody's taking notes. In any case, all the janitors can show is what they believe to be a good node title. And that's not some special skill you go to school for... we all have concepts about what makes a node title good.
I know I am suggesting a way of thinking about titling that is a bit different. To sum it up, the pool of titles chosen by throngs of people in the spur of the moment will actually be higher quality than the pool of titles chosen by a small group of people removed from the issues and acting on preconceived notions about what a good title is.
-sauoq
"My two cents aren't worth a dime.";
| [reply] |
|
I know how to spell "occurrence" correctly but I wouldn't hesitate for one second to reply to a node with "occurrance" in the title.
I don't really think that's the point. I think the point is that, if I'm looking for information about 'occurances of numbers', I will find nodes with the misspelling and participate in them. Whereas, if I look for 'occurrances of numbers' I will find nodes with the correct spelling and participate in them instead. It's all about people who are interested in the same information being able to find each other.
By ensuring the correct spelling in the node title, you can bring the two groups together. A simple Search for 'occurrance' would find all the nodes about occurrances of numbers, and a Super Search for 'occurance' would still find those with the misspelling preserved in the body. Everyone wins.
| [reply] |
|
A reply falls below the community's threshold of quality. You may see it by logging in.
| A reply falls below the community's threshold of quality. You may see it by logging in.
|
Re: Consider this: What makes a good node title?
by itub (Priest) on Nov 04, 2005 at 04:13 UTC
|
I think "program line" is a downright unarguably terrible title. It doesn't say clearly what the node is about. I either have to waste my time reading a node that I may not care about just to figure it out, or ignore it altogether (which is what I did when I saw it in the newest nodes list).
Besides 341118, I recommend Microcontent: How to Write Headlines, Page Titles, and Subject Lines for some good advice about writing titles in general. | [reply] |
|
I think "program line" is a downright unarguably terrible title. It doesn't say clearly what the node is about.
As you say, that's what you think. You're wrong on at least one point though... I'll argue it. As I said in my reply to swampyankee below, I knew what the node was about when I read the title. So, there's a little empirical evidence to counter your opinion that the title doesn't (didn't) say clearly what the node is about.
As I also said to swampyankee, the fact that you think it is a poor node title doesn't automatically indicate an absolute quality about the title. It isn't a title that works for you, but there may be many people that it does work for. In the long run, almost every question will be asked with a wide variety of titles and complete coverage will be the natural result. That is, if we don't go unnecessarily mucking with those titles.
-sauoq
"My two cents aren't worth a dime.";
| [reply] |
|
Let me qualify a bit what I said. I didn't really think it was unarguable, but I fell into the temptation of borrowing the wording from your original post. :-) Sorry about that.
I also guessed what the node might be about, so I'll agree that it wasn't impossibly obscure. But I found it a bit ambiguous, and subject to improvement.
The fact that I think it is a poor node title doesn't indicate an absolute quality about the title? Of course, I agree with that. That's exactly why we can vote on considerations: if enough people vote to "keep" the original title, it doesn't need to be changed. That way several viewpoints are taken into account.
| [reply] |
|
|
As I said in my reply to swampyankee below, I knew what the node was about when I read the title. So, there's a little empirical evidence to counter your opinion that the title doesn't (didn't) say clearly what the node is about.
This kind of "empirical evidence" is 100% meaningless.
The node could be titled "Weird Problem", and
*somebody* would step forward to say he understood
what it was about, from the title. ("I mean, just
last week I had this thing happen to me, and I thought,
Wow, what a weird problem!")
Objectively, the title "program line" is ambiguous
in the extreme. Sure, *somebody* might immediately
understand it, either by being on the same perversely
obscure wavelength as the poster, or by pure chance,
but that does not make the title clear or good.
the fact that you think it is a poor node
title doesn't automatically indicate an absolute
quality about the title
He didn't reach the conclusion that the title is
"unarguably horrible" by subjectively thinking to
himself, "Hmmm... what images does this title
bring to me, personally, in a free association
framework?" Your suggestion that anyone
who claims a title is unclear is obviously making
this claim based solely on some touchy-feely
subjective personal experience is either
deliberately obtuse or just plain poorly thought
through; in either case, from a linguistics
standpoint, it's plain wrong.
Objectively speaking, the phrase
"program line" has a large
number of possible meanings in the context of
SOPW, and for every person who reads it and
immediately gets the right idea, somebody else
will read it and immediately get the wrong idea.
No amount of irrelevant "I think, you think,
works for me, works for you" subjectivist
drivel will change that.
| [reply] |
|
Re: Consider this: What makes a good node title?
by ysth (Canon) on Nov 04, 2005 at 04:42 UTC
|
My take on this is that the title is part of the content, and shouldn't be tampered with, both for the benefit of the poster and the benefit of readers. For instance, a node titled "How do I reggex PERL" tells you a lot about the poster. However, adding additional words for searching purposes (or adding correctly spelled ones) can add a lot of value. So I'd like to see title changes just add additional info at the end, perhaps in () or []. Just my 2¢. | [reply] |
Re: Consider this: What makes a good node title?
by spiritway (Vicar) on Nov 04, 2005 at 01:13 UTC
|
Often in a dictionary, an "alternate" spelling has an entry pointing you to the correct spelling. Perhaps something like this could be applied to titles - go ahead and retitle, but keep the old one and use it as a pointer to the new one.
In general, though, I think your suggestion would be a disservice to the people you're wanting to help. Sure, we might be able to protect them from their naive mistakes by keeping the naive titles, but in the real world outside the Monastery, they'll still be running into this problem. All we'd have done is reinforced their naivete, not helped them to learn how the big boys work.
| [reply] |
|
All we'd have done is reinforced their naivete, not helped them to learn how the big boys work.
I don't even know where to start with that comment. It smacks of self-aggrandizing egotism. We are the "big boys" and it is our job to eliminate their "naivete", huh? C'mon... let's lose the drama and put things in perspective. We are here to provide help with Perl. Our purpose is no grander than that; nor need it be; nor could we fulfill it if it were.
-sauoq
"My two cents aren't worth a dime.";
| [reply] |
|
Perhaps if you take another look, you'll see what I actually wrote instead of what you say I wrote. First, I did not say we're the "big boys". I said we're the Monastery. The big boys are out there in the "real" world. Second, I did not say our job is to eliminate anyone's naivete. I suggested that we would be doing a disservice to people by fostering their naivete. There is a significant difference between the two statements.
What you have done is to distort my words to make them more extreme, and then shown that the extreme position is untenable. That is known, among other things, as a "straw man" argument. It is specious.
As for any supposed "self-aggrandizing egotism", that statement is false. I am one of the newer people here. I am brand new to Perl. I can't write even a simple script without having to check the books to see just how it's done - and when I do finally get something written, it looks suspiciously like C, because that's the language I'm coming from. I have no illusions about being a top Perl programmer, or even a competent one. I have a long way to go.
Still, despite all this, I would prefer that I not be talked down to, that the language I am asked to use here match that language that is used outside the Monastery. Yes, it's nice to know that when I get stuck, someone will likely come along and explain something to me, but I still feel it's a good thing to be allowed to flounder as I grope my way to understanding. Babying me now only delays the time when I must learn all the stuff I need to know. "Out there" there aren't likely to be so many helpful, understanding people, as here in the Monastery (present company excepted). I much prefer to learn as much as I can here, before getting knocked around in the so-called real world.
Sorry if I didn't give your idea my complete approval, but I'm entitled to my opinions, as are we all.
| [reply] |
A reply falls below the community's threshold of quality. You may see it by logging in.
|
Re: Consider this: What makes a good node title?
by Tanktalus (Canon) on Nov 04, 2005 at 15:24 UTC
|
I'm sorry, but I simply don't buy this argument, nor any of your follow on arguments in the thread. You're arguing to do a disservice to both the original poster (by not gently correcting him/her), and the newbie searcher (by reinforcing bad searches), all in the name of allowing a particular node to be found quickly. Meanwhile, we're getting in the way of more advanced searchers (it's not a stretch to think that someone who has more experience in CGI is going to need to search for help on signals or POSIX), and encouraging (by our passive acceptance) these bad subject lines.
Many of the people asking questions are also experienced. You're arguing to cater to the lowest common denominator. I would suggest instead that when someone fails to search properly, we, instead, reply and gently Help Super Search Newbies instead. That way we help them help themselves, and still have a search system that works.
| [reply] |
|
You're arguing to do a disservice to both the original poster (by not gently correcting him/her), and the newbie searcher (by reinforcing bad searches), all in the name of allowing a particular node to be found quickly.
And I don't buy this argument. As far as the original poster goes, well, correction isn't always a service whether it is provided gently or not. On the whole, people aren't coming here looking for correction on their spelling or "correction" on the way they choose to title their online queries and musings. Gratuitous corrections are unnecessary and sometimes viewed as nothing short of impolite. And, as for the newbie searcher, I rather think that requiring him to polish his searching skills before providing him with useful results is doing him more of a disservice.
Meanwhile, we're getting in the way of more advanced searchers
No. Advanced searchers would be using Super Search effectively anyway. How is it that you think retitling nodes is helping them?
You're arguing to cater to the lowest common denominator.
I don't actually agree with that characterization but, even if I did, I would deny that that's a bad thing. That approach is often just the pragmatic approach and I do take it when I feel it's appropriate. For instance, I readily admit that I restrict my lines of code to less than 80 characters to cater to the lowest common denominator...
The reason I don't agree with that characterization in this case, however, is that the word "lowest" doesn't apply and the assumption that it does apply is the root of the problem. The fact is that the best titles are not the ones you think are best. In fact, the whole concept of best in this instance is a local optimization and a function of the searcher. Retitling nodes in an effort to bring them closer to our preconceived notions of what makes an ideal title is necessarily introducing inefficiency. Good search coverage is, essentially, an emergent phenomenon. The assumption that we can help it along by retitling nodes is fallacious.
I would suggest instead that when someone fails to search properly, we, instead, reply and gently Help Super Search Newbies instead.
You are assuming that we will know that they searched first. How many people just look for nodes that answer their question and move on if they can't find one because they are too rushed or maybe just too shy to post? Besides, plenty of people post questions that have been asked and answered before. I'll argue that telling them all how to use Super Search is counterproductive because, after enough of that, all that's going to be found when someon searches is a bunch of nodes telling them how to use Super Search. In other words, if they ask, answer and occassionally mention Super Search. That's the best way to do it and that's the way it tends to be done.
-sauoq
"My two cents aren't worth a dime.";
| [reply] |
Re: Consider this: What makes a good node title?
by swampyankee (Parson) on Nov 04, 2005 at 16:30 UTC
|
I'm going to agree with ysth's itub's comment: the node title "program line" was terrible; indeed I feel it fit into sauoq's category of "downright, unarguably terrible", as it was both too generic and too non-descriptive. The title "program line" did not, even vaguely, describe the question. The replacement title was better.
I also tend to think that supersearch should (if it already isn't) be "fuzzy" enough so that common misspellings of common words (such as occurrence) should all match; my bete noire is affect|effect, which I never seem to get right. Other people have other spelling blind spots (in one company I used to work for the accepted spelling of mnemonic was pneumonic; I actually had the correct spelling flagged by one of my supervisors). There are also spelling differences by region and dialect, so using supersearch to find "Syntax colorizing tools for Perl" should (and probably does; I've only used supersearch a few times) find the large node originating in Australia about "Syntax colourising tools for Perl".
One category of node title change I would add is that node titles in the instant messager version of English should be re-titled.
Is there an English to IM translator out their? Who is up to writing Lingua::English::IM?
| [reply] |
|
I'm going to agree with ysth's comment: the node title "program line" was terrible; indeed I feel it fit into sauoq's category of "downright, unarguably terrible", as it was both too generic and to non-descriptive. The title "program line" did not, even vaguely, describe the question.
Firstly, that was itub's comment. Secondly, and more importantly, your opinion about the title is only that... an opinion. Maybe it is inexplicable, but I knew exactly what the question was about when I read the title. I suspect I wasn't the only one. So, it may have been a poor title for you but that doesn't translate to some absolute notion of a poor title.
Retitling should be used only for those cases where there really is no difference of opinion as to the quality of the title. (Titles which are found offfensive by some might be considered an exception.)
-sauoq
"My two cents aren't worth a dime.";
| [reply] |
|
I too, in my humble opinion, find "program line" to be a horrible title. It's not a question. It's a thing. It's ambigious. Congratulations if you can correctly guess what the question is going to be, but i find that there a lot of poorly chosen subject titles because it is hard to come up with a synopsis that describes the question succintly enough to fit. Most SoPW are too busy concentrating on their problem, and coming up with a good title is second to getting their question answered.
"How Do I Get the Line Number In My Error Message?" is much much better, in my humble opinion.
Even better would have been "How Do I Get the Program Line Number In My Error Message?" ;)
| [reply] |
|
|
|
|
|
|
sauoq I find your logic impeccable yet unconvincing.
This is a problem which extends beyond logic. Your arguments can be logically overturned, I don't see any value in doing
that since it doesn't address the real and valid concern you
are attempting to put forward.
Rather I suggest that Perlmonks is popular and well regarded
because of its culture. I say this culture has two core
characteristics:
-
correctness - care about doing things well in the large and small matters. Ha! note this foofaraw regarding re-titling.
-
generosity - care about others. As a group, monks kindly treat the naive, lost, and abrasive strangers.
I think the top-node demonstrates a fine sense of generosity in encouraging us to consider the preferences of others.
Applying this to something so subtle and important as habits of thought is worthwhile.
Correctness leads us to not have P3r1munks
painted on our gate. That same desire has us retitling
nodes. This thread is inspired by a marginal case: the
node could reasonably stand with its title. That said, I much prefer that the title was changed.
I feel the "broken window syndrome" is germaine. I am repeatedly impressed by the timeliness of our editors and considerers. I think the speed with which these functions
are executed is important. When a janitor sees a mess
and quickly moves to clean it up I find myself reluctant to criticize: It's too clean.
print reverse split //, "BQ gninrom yadnom rehtona tsuj";
Be well,
rir | [reply] [d/l] |
|
First, my apologies to both ysth and itub for mis-attributing the comment (They'll get messages).
Secondly, I absolutely agree that "good" vs "bad" node titles is a subjective decision. I'm disagreeing with your opinion about this particular case.
| [reply] |
Re: Consider this: What makes a good node title?
by Anonymous Monk on Nov 03, 2005 at 23:01 UTC
|
| [reply] |
Re: Consider this: What makes a good node title?
by dws (Chancellor) on Nov 03, 2005 at 22:51 UTC
|
I pretty much agree. Set the bar high on retitling. When retitling, optimize for searchability.
| [reply] |
Re: Consider this: What makes a good node title?
by phydeauxarff (Priest) on Nov 03, 2005 at 23:37 UTC
|
| [reply] |
A reply falls below the community's threshold of quality. You may see it by logging in. |
Re: Consider this: What makes a good node title?
by jZed (Prior) on Nov 04, 2005 at 15:22 UTC
|
Here's an idea which I'm not volunteering to implement (so you can safely ignore it :-)) - let's have an optional keyword field *in addition* to a title field. New users would see the field and perhaps think "oh, someone might want to search for this node". Instead of considering for retitling, picky other users could just add some keywords. | [reply] |
|
I agree with that. One root of this discussion is that some people, like me, think that the main purpose of the title is to describe the node in a clear and unambiguous way, whereas others (such as sauoq) think that its main purpose is to be a source of keywords for searching. The best solution could well be to have a list of keywords independently from the title. But I'm not volunteering to implement it either. :-)
| [reply] |
|
Isn't all this already implemented? I could have sworn there was a keyword nodelet someplace.
| [reply] |
Re: Consider this: What makes a good node title?
by ambrus (Abbot) on Nov 03, 2005 at 23:19 UTC
|
| [reply] |
Re: Consider this: What makes a good node title?
by Aristotle (Chancellor) on Nov 04, 2005 at 22:05 UTC
|
Would you agree that a minimum convention for retitling, if it has to be done, should be that it should try to preserve all the words in the original title (minus stopwords, obviously), ideally including phrasing?
Makeshifts last the longest.
| [reply] |
|
| [reply] |
Re: Consider this: What makes a good node title?
by Perl Mouse (Chaplain) on Nov 04, 2005 at 10:33 UTC
|
There's another reason not to retitle nodes to something someone already knowing the answer would use: it gives insight what the questioner knows (or doesn't know, or has a wrong ideas about).
| [reply] |
Re: Consider this: What makes a good node title?
by SamCG (Hermit) on Nov 04, 2005 at 15:58 UTC
|
I can't imagine prejudicing the title against those who spell correctly is a good thing. I also have to believe that statistically, more people will spell a word correctly (or at least realize their mistake) than use any particular misspelling.
A title like "How do you reggex PERL" does tell you something about the poster, but certainly you would learn everything it tells you on reading the post itself. I don't see that as valuable information, and it's more likely to prejudice readers against the poster before they've even read the question.
Personally, I'd probably completely ignore a post with a title like "Program Line". Since most posts are responded to quite quickly, I think the more explicative title will cause it to draw more attention. I think it's also likely to make later searches easier rather than harder for most. | [reply] |
A reply falls below the community's threshold of quality. You may see it by logging in. |
|
|