3 ms·

I built this several months ago as I was interested to see how the titles of posts evolve/change on HN. It turns out lots of titles are edited every day (in bot

I built this several months ago as I was interested to see how the titles of posts evolve/change on HN. It turns out lots of titles are edited every day (in both subjectively and objectively good and bad ways!) and I've found it interesting to see how titles have evolved. The whole thing is automated and is built around a Ruby script that pulls down the titles on the front page at a frequent interval. Any titles for specific story IDs that change get tracked and rendered out to a static HTML page hosted on Netlify. It runs by itself without incident so far. This has been on Show HN before but I was kindly invited by dang to repost it.

cjamesd 7y ago
Please tell me you plan on going meta and editing the title of this post.
- dang 7y ago
  Good idea. I put "test" up there. When will it show in the list? Edit: there it is. Title reverted now.
paulgb 7y ago
Cool! Since you have the infrastructure already, it would be neat to track URL changes as well. Probably less frequent, but I bet they happen at least daily (e.g. if a submission is flagged as blogspam.)
- petercooper 7y ago
  Yes, that would be possible, though I think it should rendered to a totally separate page to avoid confusion. I didn't realize URLs were edited here, so it was off my radar till now :)
carrozo 7y ago
Nicely done! It would be fantastic to see this embedded between post and comments on every piece here!
blondin 7y ago
love simple but useful apps like this! how's the hosting with netlify?
- petercooper 7y ago
  I'm very happy with Netlify for numerous things. In this case, Netlify is just running a build script that converts the JSON I create at http://s3.amazonaws.com/jeanfromeastenders/current.json http://s3.amazonaws.com/jeanfromeastenders/current.json into the final HTML page.
sarosh 7y ago
Any plans to release the code? Which scraper? Chrome headless?
- iudqnolq 7y ago
  Since there's no interactivity or logins or anything complex curl would probably do the job.
- petercooper 7y ago
  It's not very good, very basic and naive (but works!) https://gist.github.com/peterc/b325cdea7561528c506f073f477412fa https://gist.github.com/peterc/b325cdea7561528c506f073f47741... The actual turning the JSON into HTML takes place on Netlify. If you wish to render the JSON for your own benefit, the current version is always at http://s3.amazonaws.com/jeanfromeastenders/current.json http://s3.amazonaws.com/jeanfromeastenders/current.json
JoshMcguigan 7y ago
> Any titles for specific story IDs that change get tracked.. Can you talk more about the data infrastructure?
mft_ 7y ago
OOI, where does the Ruby script run?
- petercooper 7y ago
  Just on a random VPS I have via a cronjob. Ideally I'd run it as a Lambda job and then I wouldn't have any infrastructure at all. I just didn't get round to it after prototyping.
samstave 7y ago
This would be really good to apply to watching spin on various verticals. Such as entertainment and financial articles... Recall during the banking crisis of 2008 there was the article stating that the EU was about 16 TRILLION in debt crisis, but then that was too alarmist and they didnt. Want that getting out, so they edited the title, but they forgot to change the UrL which had the original title in it.
- Scoundreller 7y ago
  Oh god yes. I hate seeing that an article was « updated ». Well what the heck changed since I read it yesterday?!?!? And this is a well-regarded national newspaper in Canada.
- dang 7y ago
  News outlets do that all the time. NYT in particular. Often they don't change the URLs, and I've always assumed that's for technical reasons—either there's a system that's treats URLs as immutable, or it's copied through multiple systems that the people who change the title can't edit, or some such thing.
  - CPLX 7y ago
    Perhaps you’ll appreciate this: https://mobile.twitter.com/nyt_diff https://mobile.twitter.com/nyt_diff
bambax 7y ago
I just posted a very similar thing implemented as a browser extension (HNtitles[0]) so that it shows the original text in the page without any effort for the user. (I had written a first version of this some time ago[1] as a bookmaklet, but a browser extension is simpler to use.) [0] https://news.ycombinator.com/item?id=21621411 https://news.ycombinator.com/item?id=21621411 [1]: https://news.ycombinator.com/item?id=6573188 https://news.ycombinator.com/item?id=6573188
- petercooper 7y ago
  I think it works well as an extension for people who want to keep an eye on this sort of thing. I only wanted a casual, occasional look at the edits, but HNtitles looks ideal for the power user - good work.
  - bambax 7y ago
    Thanks. The list you are making is very good, it's a destination in itself. As said elsewhere, the extension helps to avoid clicking on the same story twice when the title changed. (I will add url monitoring too.)

cjamesd 7y ago

Please tell me you plan on going meta and editing the title of this post.

dang 7y ago
Good idea. I put "test" up there. When will it show in the list? Edit: there it is. Title reverted now.

paulgb 7y ago

Cool! Since you have the infrastructure already, it would be neat to track URL changes as well. Probably less frequent, but I bet they happen at least daily (e.g. if a submission is flagged as blogspam.)

petercooper 7y ago
Yes, that would be possible, though I think it should rendered to a totally separate page to avoid confusion. I didn't realize URLs were edited here, so it was off my radar till now :)

carrozo 7y ago

Nicely done! It would be fantastic to see this embedded between post and comments on every piece here!

blondin 7y ago

love simple but useful apps like this! how's the hosting with netlify?

petercooper 7y ago
I'm very happy with Netlify for numerous things. In this case, Netlify is just running a build script that converts the JSON I create at http://s3.amazonaws.com/jeanfromeastenders/current.json http://s3.amazonaws.com/jeanfromeastenders/current.json into the final HTML page.

sarosh 7y ago

Any plans to release the code? Which scraper? Chrome headless?

iudqnolq 7y ago
Since there's no interactivity or logins or anything complex curl would probably do the job.
petercooper 7y ago
It's not very good, very basic and naive (but works!) https://gist.github.com/peterc/b325cdea7561528c506f073f477412fa https://gist.github.com/peterc/b325cdea7561528c506f073f47741... The actual turning the JSON into HTML takes place on Netlify. If you wish to render the JSON for your own benefit, the current version is always at http://s3.amazonaws.com/jeanfromeastenders/current.json http://s3.amazonaws.com/jeanfromeastenders/current.json

JoshMcguigan 7y ago

> Any titles for specific story IDs that change get tracked.. Can you talk more about the data infrastructure?

mft_ 7y ago

OOI, where does the Ruby script run?

petercooper 7y ago
Just on a random VPS I have via a cronjob. Ideally I'd run it as a Lambda job and then I wouldn't have any infrastructure at all. I just didn't get round to it after prototyping.

samstave 7y ago

This would be really good to apply to watching spin on various verticals. Such as entertainment and financial articles... Recall during the banking crisis of 2008 there was the article stating that the EU was about 16 TRILLION in debt crisis, but then that was too alarmist and they didnt. Want that getting out, so they edited the title, but they forgot to change the UrL which had the original title in it.

Scoundreller 7y ago
Oh god yes. I hate seeing that an article was « updated ». Well what the heck changed since I read it yesterday?!?!? And this is a well-regarded national newspaper in Canada.
dang 7y ago
News outlets do that all the time. NYT in particular. Often they don't change the URLs, and I've always assumed that's for technical reasons—either there's a system that's treats URLs as immutable, or it's copied through multiple systems that the people who change the title can't edit, or some such thing.
- CPLX 7y ago
  Perhaps you’ll appreciate this: https://mobile.twitter.com/nyt_diff https://mobile.twitter.com/nyt_diff

bambax 7y ago

I just posted a very similar thing implemented as a browser extension (HNtitles[0]) so that it shows the original text in the page without any effort for the user. (I had written a first version of this some time ago[1] as a bookmaklet, but a browser extension is simpler to use.) [0] https://news.ycombinator.com/item?id=21621411 https://news.ycombinator.com/item?id=21621411 [1]: https://news.ycombinator.com/item?id=6573188 https://news.ycombinator.com/item?id=6573188

petercooper 7y ago
I think it works well as an extension for people who want to keep an eye on this sort of thing. I only wanted a casual, occasional look at the edits, but HNtitles looks ideal for the power user - good work.
- bambax 7y ago
  Thanks. The list you are making is very good, it's a destination in itself. As said elsewhere, the extension helps to avoid clicking on the same story twice when the title changed. (I will add url monitoring too.)