Why we revert to original titles

438 points by pg ↗ HN
In the HN guidelines, we ask submitters to use original titles when possible. When they don't, we often change the title of a post back to the original title of the article.

There is an ongoing trickle of complaints about this, as if we were engaged in some sort of sinister conspiracy.

Titles on HN are not self-expression the way comments are. Titles are common property. The person who happens to submit something first shouldn't thereby get the right to choose the title for everyone else. This would be clearer if we didn't let submitters enter a title-- if our software simply let people submit urls, and retrieved the title from the page. We don't do this because it's too inflexible. Some articles have titles that are too long. In others the subtitle makes a better title. But the fact that a title field is editable doesn't make it comment.

It's true that when submitters change titles, their new titles often contain more information than the article's original title. But a significant percentage of the extra information added in this way is false. The only way we can tell if a newly created title is accurate is to read the article, and we're not about to read every article submitted to HN. The only option is to revert to the original title, which is at least what the author intended.

(We do sometimes change titles from the original when the original title is egregious linkbait, or false. We have also, since the beginning when our users were largely YC alumni, put e.g. (YC S13) after the names of YC companies in titles. But these are not the types of changes users mean when they complain about moderators changing titles.)

If we had infinite attention to spend on moderation, we could read every article and decide whether each user-created title was better than the original title. But we don't. Moderating HN is no one's full time job. And frankly it's not that big a deal anyway. If we're going to expend cycles trying to fix something about HN, the increasing prevalence of mean and stupid comments has a much higher priority than the fact that authors' original titles are not maximally informative.

230 comments

[ 2.9 ms ] story [ 248 ms ] thread
One of the most helpful title additions is the (YEAR), for older articles. This gives nice context for the reader, before clicking the article.
Yes, an obvious cue for something that is not news, but may be interesting seems wholly appreciate. I think HN is pretty good about this in general for articles not in the current calendar year. Such a modification is not editorializing or biasing the content in anyway, too. So this keeps a consistency in place on that front.
You could make everyone happy by adding a comment field to the submission form so that the submitter could add their own sub-title.
I don't understand why HN currently considers link submissions and text submissions to be mutually exclusive. Most of the problems with titles seem to stem from the title being overloaded to carry additional information. A text field for link submissions would relieve these problems by providing a proper place for additional information that the submitter would like to add.
From the FAQ:

How do I make a link in a question?

You can't. (This is to prevent people from using this method as a way of submitting a link, but with their comments in a privileged position at the top of the page. If you want to submit a link with comments, just submit it, then add a regular comment.)

He just explained why. Article postings are community property. The first person to post an article wasn't intended to have a privilege to editorialize or summarize the article.
You can always post a comment on the article submitted which indicates the reasoning you had for submitting it. When you submit it you end up on the 'new' page with your submission on top, if its a duplicate you end up on the comment page for the original submission.
You can always post a comment on the article submitted which indicates the reasoning you had for submitting it.

Then every bland default post title requires a visit to the post page to read the extra detail.

Not completely unreasonable, but not an good UX pattern either.

Gives me an idea for a Greasemonkey script I'll never actually write: fetch the first comment for a post and make it visible when mousing over the title on the list of posted stories.

I think that's covered by the issue of titles being common, not personal. If two people both submit the same story, the first person to submit it should not get extra editorial privileges over later submitters.

I suppose you could have a separate "subtitle" field that anyone can add to, and vote the best/worst like comments, but that seems like more trouble than it's worth.

Thank you for clearing this up.

Has it been considered having a subtitle showing [previously titled: xxx] or some such when a title is edited?

Or possibly relying on a flagging feature along the lines of "misleading or editorialized title"? Rather than just changing all/most titles?

I think these are exceptionally good ideas.

The only reason I typically care about this conversation is that a year ago or so I was tracking an article that I saw the "new" page, and it got its title changed to something insanely literal like "Post #4". Which of course tanked it's chance of getting off the new page and me learning anything about the topic.

Either of your suggestions would have solved this problem.

Having a "flag this title" feature is the best suggestion I've seen so far. If there aren't enough moderators to read articles before reverting titles, then having the community do it makes so much more sense than doing it blindly.

This approach gives intelligent first posters the necessary flexibility to change the title for any of many valid reasons (I still can't believe your comment that they changed the title of that serial killer article, killing one of the most chillingly powerful articles of the last month), yet it prevents first posters from abusing their priority even more effectively than the current approach.

If you're worried about title originality, why don't you write a scraper instead of asking for people to submit urls manually? Would be more efficient. But, If you care about real people posting stuff, then they should be allowed submit whatever they see appropriate. To mitigate the frustration, I think you could allow urls to be posted multiple times with a specific scientifically backed grace period of "n-hours" until no more duplicates are allowed. People can still up-vote the titles they like best and the winner takes all, making the other titles appear like collapsed sub-captions below the main-title. Hope you understand what I mean. This would solve two problems at once. A) Custom title's, no censorship. B) Valuable data about winning-titles, that can be used to train a stochastic model to predict the best titles. How you use the data from B) is up to you.

I mean people aren't stupid enough to change the Title of the "Higgs Boson" to "Bananas". Sorry, if this comes over wrong. I respect you and this is just critics on your software's policies.

I don't complain about changing titles, but now that we're here: I don't mind them all too much, except when the submission comments are tied to the title some way (e.g., a submission's title reflects a commentary on a small part of a larger article). If the submission title were to change to match the article's title, the context of the comments are lost.

Maybe keep the submission title as a title attribute on the href? I have no idea what that does to SEO (or if anyone cares).

(comment deleted)
This might've been covered elsewhere, but - with regards to increasing meanness and stupidity on the site: have you considered just adding a line of text under the comment field that says something like "reminder: don't say things that you wouldn't say to a person's face in real life"?
I wish he would just update the site guidelines. Some of the meanest, stupidest comments on the site can be rationalized as fitting into the wiggle room of the guidelines.
But the only people who read the guidelines are those who intend to follow them.
Many of the titles that I wish were re-phrased were titles from mainstream news outlets. Good original content tends to be headed with good titles.
> we're not about to read every article submitted to HN.

How many articles reach the front page per day? I'd imagine it is a small percentage of the total

Still enough material that nobody smart enough to be involved with HN/YC would want the task of reading the articles just to verify whether the altered title is suitable (remember, the task is to verify the alteration is fair & accurate, not whether you like or agree with it). Would you want that job?
But that's greatly influenced by titles; a classic chicken & egg problem.

scenario 1: content w/editorialized [sensational] title -> front page -> title correction -> content on front page still

scenario 2: content w/author's [uninformative] title -> never reaches front page -> never gets reviewed

Scenario 3: content with (author's/publisher's) [sensational] title -> never reaches front page because it sounds like a bunch of BS
I agree that this policy is sensible from a process standpoint. However, it does end up penalizing the writers who are bad at SEO, or just don't care...and at the same time, it rewards the sites that do cynically partake in link bait titles, all the while being little more than blogspam.

I think my problem is that when a headline is clearly too vague and someone adds a non adjectivey headline, the mods go out of their way to revert it, doing a disservice to everyone. If monitoring titles is a burden, then it seems like it'd be less work in these cases to leave the clarified titles...the community is usually good about flagging it.

Also, do HN mods revert to headline or the title tag? That is, can submitters choose from either (this is significant for most New Yorker articles, which have very short heds by properly descriptive title tags)

Yeah, my main problem with the current policy is that it lowers the overall quality of content by encouraging content from places that are optimizing their titles. Some places even write titles deliberately targeted at HN. Stuff not intended to "go viral" or be marketed, but rather just to provide useful information in a non-HN context, often ends up with completely unintelligible, out-of-context titles, like "Update". But that material is often actually better.

I guess I could try to adopt a personal policy of only reading submissions that have vague, out-of-context titles, and see if it works as a kind of reverse heuristic.

> the increasing prevalence of mean and stupid comments has a much higher priority than the fact that authors' original titles are not maximally informative.

Give a small number of trusted users a "mega downvote" - it takes a comment to -2 with a single click.

(Just for clarity: I don't want this button. I'd be a terrible person to give it to.)

AKA how to ensure the hivemind remains strong
It's nothing to do with hivemind - it's entirely to stop mean and stupid comments.
Totally agree. Most people here are smart, we don't need submitters editorializing, which is usually what title changes are.
Being smart, we can also generally tell when a title is editorializing, and weigh that against the quality of the original source in deciding whether or not to click. If all titles are auto-reverted, then we're at the mercy of the crappy editorializing of the publisher; I feel I have to check out stuff with stupid titles because maybe there's a cogent article behind it.
I can't tell if a submitter is editorializing or the original article is until I click through. If I know it's the publisher because titles aren't altered, I can simply not click through if I'm not interested. If it's the submitter, then maybe there's a great article there that I'd miss because of it.
I disagree. Most editorializing is obvious, because people who can't see the difference between facts and personal opinion tend to make free with the adjectives.
I think you misunderstood. It's obvious when there is editorializing. However without clicking through, there is no way to know if it's the article's actual title doing it, or the submitter, because you don't know if it was changed. Unless all submissions use the actual title (or none I suppose).
Ah, I see. I feel like I am pretty good at guessing based on the source and the poster history, but I'd have to try a controlled test to know that's not just my bias.
>This would be clearer if we didn't let submitters enter a title-- if our software simply let people submit urls, and retrieved the title from the page. We don't do this because it's too inflexible. Some articles have titles that are too long. In others the subtitle makes a better title.

The problem is you've created a horrible half way house. There is a class of submissions that only make sense or attract interest with a custom title. These generally get reverted to some meaningless title which then prompts a lot of pointless discussion about the title change. If you don't have the man power to review custom titles, and don't trust the community to do it then disallow them other than in the case of manually editing down titles that are too long. It means missing out on a certain class of submissions, but those are mostly a mess these days anyway because they get filled with people talking about the automatic title change and people confused about why the link was submitted and upvoted.

> a horrible half way house

Horrible might be a little strong. "Mildly confusing," or, "suboptimal in certain respects" are more accurate.

I expect you to be shadow banned without reason in 3..2..1...

:D

Hahahaha the part time site admins running this place for free will totally just shadow ban you for saying something they disagree about, funny joke.
Yeah right, there are no incidents of people getting shadow banned due to personal vengeance/ stuff they disagree about, on Hackernews. /s

And 'running it free' is not like they are doing us a favour like you make it out to be, but running it right, is.

And we all know you are one of the good admins here too :) /s

It's not like anyone would know if they did.
Happened to me, more than once.
The problem is that they have a simple policy that makes total sense based on their principles (first person to post a story doesn't have special privileges, and posts are community property) and their limitations (part time admins).

Unfortunately, that policy engages with an issue that nerds can debate endlessly --- which title is better? What constitutes editorializing? Are original author titles the optimal titles? Oh, look, there's that word "optimal" --- let's spend seventeen weeks debating it!

Therefore, it seems like there's something for us to discuss. But there really isn't a discussion to be had here. People who want titles to be managed should start their own HN alternatives. We could use more of them. Or, even simpler: if you have a story with a bad title and a new title you feel strongly about, instead of submitting and then writing a long comment, write a blog post about the story and submit that instead.

Of course, those two suggestions are much less fun to talk about than a debate about titling stories.

"write a blog post about the story and submit that instead."

A long comment and a link are probably OK, but if you just want to change the title, you'd probably get accused of blogspam.

I don't know if tpacek meant this instead, but how about writing an article briefly explaining why your title is better and then provide a follow on link to the article.

This has the advantage of getting the right title, getting out a link to the article, and having a quick synopsis of why the title change was made in the first place.

write a blog post about the story and submit that instead

Well, actually, the guidelines ask us to submit the original. But you can submit the original and link to your blog in the comments.

But that does not fix the stupid title problem the person wanted to fix in first place.
Indeed. You can't fix stupid titles. But if there is a real issue with the title- it's misleading, sexist, or what-have-you, don't fix the title. Write your opinions as a comment or a separate blog post.

The point is- the title isn't important enough to worry about- content is.

But unless you read every article that hits the front page, titles are important.

Titles are how I decide which articles to read, and often the original title doesn't explain why the article should/might be considered interesting by the HN audience.

> The problem is that they have a simple policy that makes total sense based on their principles (first person to post a story doesn't have special privileges, and posts are community property) and their limitations (part time admins).

Actually, the problem is that they have a recently changed[1] policy that makes no sense whatsoever given the limitations on this site with relation to the nature of media that's being posted. I.e. the fact that submitters cannot post a title and a lede, while the lede on a linked site may add much needed context to an otherwise ambiguous title.

Unfortunately, your head is so far up pg's ass that you were unable to realize this simple fact.

Therefore, you may continue massaging pg's balls.

[1] As recently as last year, this was the policy: "You can make up a new title if you want, but if you put gratuitous editorial spin on it, the editors may rewrite it."

HN has been around for a while, so many long-time users probably aren't even aware of the change.

username485631 (if that is your real name), nice use of the irony in drawing attention to the "mean and stupid comments" point.
> Or, even simpler: if you have a story with a bad title and a new title you feel strongly about, instead of submitting and then writing a long comment, write a blog post about the story and submit that instead.

As stated elsewhere in this thread by others in replies to your suggestion (but sadly much lower on this page, so people will continue to see this earlier comment of yours): short blog posts telling people about another post are generally considered "blogspam" and so thereby posting anything but the original article is not just widely disliked on this site but explicitly violates the guidelines.

"Please submit the original source. If a blog post reports on something they found on another site, submit the latter.".

We're already talking about a case where users think that they're smarter than the guidelines, and are choosing to ignore the guideline to use the original title.

Between submitting retitled blog spam with commentary and editorializing the title of the original source, I believe the former is a lesser evil.

Users can upvote the original source when it is submitted with the original title, while the blog spam falls off of newest. (In the unlikely case that the blog spam has a better title, it will get upvotes as you'd normally expect.)

Right now, this editorializing of titles puts a lot of burden on the moderators, and I think that any solution to this issue that we come up with that has to deal with the finite bandwidth of those moderators.

hey why not have 2 titles, original and secondary... Moderators would simply identify and delete bad secondary titles and prevent redditization...example:

A Farewell --- Android head quits

(comment deleted)
So, I continue to read tptacek's comment to mean "here is something legitimate you can do", not "here's something else you shouldn't do"; if it is the latter, as you are stating, I'm not certain why it was brought up in the first place.

Regardless, I take issue with your contention and seemingly fundamental assumption that the article's original title is somehow better and will win out against alternative titles. The entire reason people get angry about this is that the alternative titles are often much more useful: they are more precise, and they are even sometimes less sensationalized.

Sometimes titles just need to be different when the material is being viewed in a different context. I've actually been chewed out before for "editorializing the title" when I wrote the original article in a case where my original didn't even have a title because the medium (Google+) doesn't use them. In other cases, the in-context title might need to be very short, but on Hacker News it should be longer.

The places you thereby see users actually complaining about this policy are not the places where the submitted title sucked. To demonstrate just how dismissive it is to claim that that is the situation we are discussing: it's like trying to end a debate on whether airport security should strip search anyone who talks about security issues because clearly everyone who does so is a dangerous terrorist.

In fact, even people (such as myself) who complain about title changes made on Hacker News often call for title changes, as to many of us the problem is "inaccuracy", not "originality": if the title says something that doesn't seem to be true (as in the case just yesterday regarding the Tesla S's usage of rare earth metals) it isn't the person submitting the link that is "editorializing", it is the original author.

Another interesting circumstance was one of the original cases from about a year ago that made this turn into a big public argument: the visualization on the New York Times about some non-hacker aspect of politics. The visualization was insanely impressive, and someone wanted to submit the visualization. The title was clear, factual, and even quite concise: it just labeled the page as being a visualization.

In this case, there were tons of interesting comments about the visualization, and people were really excited to see it: the link was clearly a success, it was voted to the front page, and not a single person was bothered that it was there... until some moderator changed the title to the original title of the visualization, which was about some political issue, and now the comments changed in tone.

Now, people were arguing about how this wasn't worthy of being on Hacker News: that political issues should be discussed elsewhere; then a fight breaks out about the underlying politics. People are infuriated that somehow this link ended up on the front page with an insane number of votes, and no one has any context due to the policy of changing the titles of posts, even after many hours of discussion in hundreds of comments, without any historical notice that it happened.

This is the kind of situation that I think it is reasonable to be bothered by; sure, there are some people who believe that the titles should never be changed under any circumstance, and to the extent to which they exist it is convenient to lump them together with everyone else because "well, people editorialize and there's linkbait so the titles should be changed QED"; but that's not the argument that I've actually seen about title changes: people just want more nuance and more transparency.

I'm trying to acknowledge that a strict interpretation of the guidelines would lead one to conclude that tptacek's suggestion is "something else you shouldn't do", since it would be "blog spam". I also think that a well-informed title and comment giving context for the original source should be "something legitimate you can do" if it provided good value above and beyond the original source itself.

I can totally see someone submitting a HN page talking about the visualization in the NYT to some meta-HN. (I was going to remark about reddit.com submissions, but most of the top ones seem to be submissions where reddit.com was the original source.)

I suspect that "more transparency" won't happen if it results in more work for the moderators; I'd vastly prefer solutions which are self-reinforcing and can come about by the community working together within the existing framework of HN.

A lot of the value of "more transparency" just comes from "if the title is changed, make certain the old title is also visible from the comments page, so people don't freak out". One time someone submitted a link to a YouTube video of a panel discussion involving Linus Torvalds where he explicitly gave NVIDIA both the finger and the F-word. The link was a link directly to the paragraph in the video where it happened, and the title was clear that that is what was being linked.

Upvoted to the front page, holding there for hours, people were having a discussion about the appropriateness of swearing in this context, the extent to which it was a joke, and about how NVIDIA has treated the Linux community. And, of course, then the title changed: it got reverted to something like "Random Panel from Random Conference".

Now, everyone was really confused: the comments were all concentrating on this one moment in an hour long video (linked to semi-directly, but still like 30 seconds after the link), and freaking out about how people were so uptight and shouldn't be concentrating on that, and that there was all of this other content there.

When the title changes, it drastically changes the character of the discussion: while this website is also a link aggregator, I know of very few people who believe that that is its value; instead, people like Hacker News because of the comments: it is a discussion forum. You can't just change the title of the discussion forum while people are actively discussing things without causing a catastrophic situation.

This is thereby what I mean by "transparency": I don't mean a moderator spending even an extra 30 seconds explaining "why" something happened... I just mean that the site needs to make it clear that something did happen, or these massive arguments break out and a bunch of people are left unhappy and angry. As long as title changes are going to happen to discussion threads that already have hundreds of comments, you at least need to have a boolean "title changed" notice.

That aside, the "nuance" aspect (which I think is even more important than the "transparency" part, as you might just decide "old posts should never have their title changed" and the transparency is less important) would probably require more time from the moderators. I am not certain how to get around that: I think moderation fundamentally requires nuance.

edited my comment

There just need to be enough watchdogs.

A very simple solution would be that titles can be flagged by the community. Example below:

  submission X by kdzsb 22 minutes ago | flag title | flag discussion | 9 comments
Moderators would simply edit flagged bad titles. Rather than having them to watch all titles (old rules) or no titles at all (new rules).
> Yes, more power to the community. Wikipedia is still alive despite its massive openness. There just need to be enough trusted watchdogs.

Wikipedia has enormous meta community, and it's likely that such a huge amount of meta is toxic.

(comment deleted)
Minor point: flag discussion should probably be flag submission. The current flag action is more for flagging a submission not suitable for HN rather than flagging the discussion around it. Those two things are different.
Yes, that is minor but still important. Having 3 options: flag submission and flag discussion and flag title might be too much.

That's up to the admins to decide what needs more often to be managed, the submission or the discussion. Or it could be named:

    flag title | flag submission/discussion
Agreed. It creates a perception gap.
Reminds me a bit of terrible zero-tolerance rules that you see in schools nowadays. It's easier to enforce a "simple policy" with no exceptions than to make judgements, but that doesn't mean it's better.
There are all sorts of things it would be "nicer" to have Paul Graham doing for me for free. Writing fuzzers. Making coffee. Doing laundry.
This re-titling, if I understand correctly, is already a manual process, right?

So this is less "Hey PG, could you wash my dishes and do my laundry" and more "Hey PG, could you please stop loading up my dishwasher with dirty socks?" The labor is already being done, but at least one of the people doing it seems to be doing it in cases where it is detrimental.

Sure, let's go down the analogy rabbit-hole.

It's more like PG agreeing to do your laundry, but under the condition that he will put all colors together, regardless of your wishes, instructions or special sorting of laundry before you give it to him. He's willing to do some work for you, just not as much as you would like. It may even seem like he does more work, when clothes that were pre-sorted into single load sized chunks end up mixed together, but that's his process. If you want sorted laundry, use a different service.

I guess what I should be saying is that this laundry service is ripe for disruption by a more flexible newcomer. But I wouldn't say that, because that would be a low, cheap joke that would prevent me from looking too long at myself in the mirror.

To be clear, my issue is not the HN policy (as far as I understand it). I am 100% A-Okay with HN moderators reverting editorialized titles.

My issue is with the HN moderator[s] who are not using common sense while executing this policy. When an informative title gets changed to an information-free title, it is detrimental to the site. The point of using moderators to do this instead of auto-reverting all titles is so that the moderators can use some judgement. That is great, so long as they actually do that.

There is nobody doing this work for HN that doesn't deserve the benefit of the doubt, and then some. HN doesn't have a lot of moderators; it is intimately intertwined with YC the firm (something I've pointed out on HN before). Moderating it is a tricky problem.
I certainly don't mean to imply that any of the moderators have malicious intentions.
This is a great analogy, I will be stealing it and using it regularly. Seriously, well said.
This is an unfair comment. You're essentially turning a healthy, non-personal discussion on how best to evolve HN submission policies into one where its the dissenters vs. PG.

Worse, you then tack on hypothetical (and frankly ridiculous) extensions like making coffee or doing laundry (both are personal tasks with no relevance to the HN community).

I'd like to believe the community is mature enough to debate this issue without having to be subtly threatened with PG's authority.

Holy shit! Now look at this thread! The whole point is, this isn't an important enough issue to worry about. We all have real work to do. And we have people writing their asses off to a reply to a reply...
But the point of getting titles right is to save us from wasting time reading stuff we don't want to when we could be wasting/spending time reading stuff we'd rather.

To my mind a truncated page title taken from the page itself along with a subtitle provided by the submitter and optionally a user selected number of lines from the submission content itself would provided a more useful and informative page for triaging content one wishes to consume/discuss.

However it seems that moderation of titles isn't working so simply reverting to using the given page title created by the content creator is probably the best immediate option.

IMO a site like HN should be highly customisable. If I want page titles and you want the submitters titles and someone else wants only the head lines from the submitters content then why not. If you want votes hidden and I want them visible, why not. If you want votes weighted by longevity of the user and I want votes weighted by people I've upvoted, why not.

But that's not HN and primarily it's the corpus of people here rather than the site that generate the worth.

(comment deleted)
Make a blog post with your custom title, link to the article you were going to submit, add your commentary even more fully, submit your blog post.
People hate that and often call it a link-bait or the top comment ends being "Link to the Original".

But then the top comment on such submissions end up being [Original]: <link>

So it just might work.

I find people who vehemently debate relatively esoteric HN policies like title changes a little odd. In order to function, every ship needs a captain. PG is the captain of this particular ship. If you don't like a policy, you can suggest an improvement and see if he decides if it's worthwhile to make a change. The HN people are always very responsive and thoughtful, so you can't fault them for ignoring their users or making ad-hoc decisions. If you happen not to like the decisions of the captain of this particular ship, you're always welcome to get off at the next port. It's pretty simple, so I really don't understand where the complaints are coming from.
It's not that hard to understand, really. People get passionate about the things they like. They want to see those things be as successful as they possibly can. When they see a decision made that they think will hinder that, they get cranky. But they're only cranky because they care; nobody complains about the policies at a place they never go to, or gripes when a band they don't listen to changes its sound.

In that sense it's a compliment to PG and the admins that people get so worked up about things like this; it shows they've built something people care about.

The easy way to solve this problem would be to add a sub-title field on the submission page. For many submission it could be left blank, but for the ones where the submitter wants to draw attention to something special or non-obvious, like a bug in the submitted page, he could fill out the sub-title.
It just isn't that big of a deal. Either way.
Agree, this comes down to trust. Trust trust trust. If you don't have infinite time for moderating, paul, open it up to others. In fact I'd love to see a system where people can also vote on titles. Most people (aside from the OP and submitters) hate sensationalized titles and would vote up a better one. I guess Reddit kind of does this by allowing multiple submissions of the same link, but you still get a lot of garbage titles voted up anyway, and by the time the submission makes it to the front page, the title is fixed.
When I view HM news page my screen is 90% empty space. Why not add an option to display both original title and author title? This way everyone could set their preference (Original only/Custom only/Both - with additional option of which one to display first).

When a person pastes URL into box, how hard can it be to follow that URL and parse original title? This way people won't even have to input more fields.

Another option involving scraping, but allowing some human intervention, would be to have a default title - and have a new, and optional "edit the title" page for submitters.

This would hopefully decrease the number of edited titles as they would be more effort to enter - and would let the mods revert them with the click of a button as they could compare titles without reading articles.

I think the biggest problem in reverting to original titles is that oftentimes, the original title is not bad, but it only makes sense in the context of the original blog in which it appears. In a social aggregator, it suddenly doesn't make sense anymore.

Consider this title:

    A New Beginning
In the context of the PHP blog, it might indicate a change of direction of the project, a change of leadership, etc. It's a decently sensible title. On a social aggregator like HN, it is much less useful, even if printed next to a small (php.net).

We'd be better off if we let the submitted change it to:

    PHP project changes direction, elects new leader
Well, doesn't the domain affix (php.net) provide the context?
Not necessarily, for example medium blogs just show 'medium.com' with no relation to the actual blog.
php.net, sure, but what about more obscure domains?
How about this one from the typescript blog? The domain is not enough to provide the context here.

    *Announcing 0.9.1* (blogs.msdn.com)
To play the devil's advocate, we already have the URL so the title (at least directly on https://news.ycombinator.com, experience on hn-android and other clients will vary):

    A New Beginning (us.php.net) 
Is that not good enough?
No. Titles like that are often heavily biased towards regular readers of a blog/forum/interest group, and meaningless to people outside that small circle, even though the article itself may be well worth knowing about.
Github submissions are typically terrible with it. Especially because people use their personal repos as a blog of sorts, which means that it's hard to tell when a submission from Github is an official blog post, a code repo, or an article to read.

For example, https://news.ycombinator.com/item?id=6573101 is on the front page right now.

The title is:

Modeling your App's User Session (github.com)

It's a blog post by Github staff about how they're now tracking user sessions with Rails, cookies, and their DB. Unfortunately, due to the way that the Github site is structured, from the title, one might think that this is a link to a library repo for user session tracking on a mobile app. A small change to the title, like "Github's modeling of user sessions" would convey the meaning of the blog post much better than the title.

That's a good point. I've had a similar problem with submissions from google.com, which may be official google stuff or may be comment from Groups or G+ (though I think this might have been fixed by showing more of the URL, haven't checked lately).
In that case, yes. But what if the domain is somedudesname.com and that is the blog of some company's CEO? It'd be preferable to have a title like "XCompany changes direction, elects new leader".
But he responded to a specific example. If you want to come up with more examples, fine. But don't tell him he's wrong or do a "yes, but".
The example was only there to show that there are cases where editing the title helps. If you want to get to the best answer, don't argue against what someone said, argue against what they should have said.

Being right or wrong about dismissing a bad example doesn't matter. The important thing is to fix the example and then evaluate it.

Then say something like: "True, for that example. But a better example is...".

I guess I have a problem with jumping on someone who was correct.

Thank you for your help. I agree that the information in the URL is not always enough. I am glad it started a discussion and I hope pg reads the rebuttals to what I wrote. (github.com) is a good example. It shows why github was so eager to move gh-pages to github.io

Perhaps a way to get around the ambiguity (at least in case of github) would be to encourage people to use/link to gh-pages (so we see github.io as distinct from github.com) when they are available?

That's still not enough useful info. It tells us that someone has posted a webpage about A New Beginning on php.net, but it doesn't give us enough info to infer that it's a blog post about a new leader taking over PHP. Even worse if it's from some generic URL (blogspot.com, etc).
I've come across so many examples where a perfectly valid (imho) title has been changed to something which is basically meaningless with out the context.

Here's a recent one from memory:

https://news.ycombinator.com/item?id=6492781

The originally supplied title was something similar to: "My afternoon with a serial killer". It was changed to the "Center of the Universe" title.

Clicking a link titled "Center of the Universe" one would rationally think they were about to read something by an astronomer regarding the latest hubble deep space image.

The biggest problem with this is the unavoidable editorializing. Its as likely to become

   PHP project takes awful turn, elects incompetent new leader.
I would argue that is a lesser problem than nobody who is interested in PHP clicking through to the HN discussion because they don't have any indication that the post is about something they are interested in.

Anyway, re-titling horribly editorialized posts is fine. The problem is that the person/people doing it seem to exercise little to no discretion when doing it.

That's not unavoidable, and when I see editorializing like that it's very obvious that it's axe-grindey editorializing which I can flag and/or comment upon.
I think that's very potentially worse. "Changing direction" is far too opinionated.
No it is not worse! Even if the title is totally biased and link-baity and whatever at least I can still guess the approximate topic the article is about. Based on this I can then decide if i am interested in it or not. But with the original title this is often impossible.

For example a few days ago there was a submission with the title "Beer". Since i am interested in beer I clicked on it. But instead of something interesting about beer I got a piece of alcohol fueled drama in the Ruby community which I am totally not interested in.

By leaving the original title I almost always have to click on all articles on the frontpage to find the 4 or 5 articles I am interested in.

Perhaps we should get rid of the complete title to make it as unbiased as possible? Just put the number 1 to 30 on a page and link them to the articles and hide the domain via javascript. This way you prevent a lot of problems... For me almost nothing would change because I already have to click on almost all the links just to figure out what's in the article and if i am interested in it.

I would prefer simply prefixing such simple titles with the missing context, e.x.

    PHP Blog: A New Beginning
(comment deleted)
Why is there a length limit?
The semantics of using a title kind of dictate that.
The thing is that even if the original title is what "the author intended" on a blog where (a) it will usually be accompanied by some or all of the text and (b) context about who the author is is evident from the rest of the site, in my opinion many of them are essentially meaningless out of context - anecdotally, especially for more personal posts where a descriptive title or anything that seems like SEO might seem too formal. When posts are modified to these titles on the HN front page, readers are left to click either due to domain recognition (which isn't always there) or mere curiosity, without a clue what they'll find at the link. This is unfair, since the post may be highly interesting yet has to compete with many other posts with better titles. Not that big a deal, but when you're actively going out reverting titles of popular posts, IMHO, it would be better to add some basic context if easily available.
(comment deleted)
Why not just mention that on the submit page? I had no idea this was a rule. The submit page doesn't say anything about using the original title.
Or fetch the title, pre-populate the submission form with it, and then upon submit, require a confirmation along the lines of "You've changed the title of this submission. Please note that in general, we want submissions to use the original article's title except in cases where it is necessary to edit, for e.g. reasons of clarification or space. We do not allow editorializing in titles, so if this title edit isn't good, your change will be reverted by a moderator and it will not reflect well on your karma score. Are you sure you want to submit with the edited title?"
This would be IMHO the best solution UX wise.
That's a good idea.
If you're doing that, please also revisit the guideline about not posting with linkbait titles - because that's usually why I edit, and it's frustrating to replace a linkabity title with a more factual one only to see it reverted later.
At the risk of over-thinking, it might be useful to have a "rationale for edit" field that lets a submitter briefly make their case for changing the title ("orig title was linkbait" etc) to make moderators' jobs easier.
Is it possible to increase the title length, or add how many characters over the current title is. I keep copying submissions into a text editor to edit the title and it would probably easier for folks if the rejection form told how many characters needed to be deleted.
Allow the title field to be blank, in which case the title is auto-fetched. (I just tried leaving the title blank and it barfed at me).

If a custom title is submitted, require an explanation. Possibly display both the custom title and the original title.

Hey pg, I've often wondered if something similar to this might help with the civility issue.

I'm not thinking blink tag gaudy :) - but seeing an "unmissable" reminder message as you are typing your comment into the textarea would probably be helpful in curbing abuses.

Another option would be to treat different titles as different entries, and let the better-titled links get voted to the top. But the current system seems to already work pretty well. If it means more effort is expended in moderation, I submit that that effort is probably not a waste. Better editing makes better reading.
This makes plenty of sense.

Either stick with this strategy or do what Digg.com is doing these days; a main title and the little subtext thing that they almost always use for a one-liner joke, but sometimes for serious commentary/secondary-title.