Hacker News Reader

Comment by imurray | Hacker News Reader

Can anyone get the original text?

It is a bit annoying how the BBC non-transparently correct their articles. The old version is cached here: http://www.webcitation.org/60evsCAry

View on HN

We run a web content monitoring service, every media organisation changes their articles as they get new info. No reference is ever made to these changes. It's actually really interesting seeing how stories get updated.

I'm sorry, but there's a difference between changing your article for updates and editor required corrections, and changing your article to look as if you never bit the bullet. BBC did exactly the latter, they didn't put an "Update: It has been determined this was a hoax" or "Please forgive our error, we have made a mistake..." anything.

Reading the cached article, and then going back to the original, I think it is fully within your duty to inform your readers that the article has been updated specifically for the reason of clarification and acknowledging a mistake was made, and that new information has come to the fore.

Trying to hide behind that by virtually rewriting the article is just unsavory, and I expect better from the BBC News.

Sorry if it seems like I'm taking a high road here, but I am; retroactive continuity has NO place in journalism.

You're right, and I agree with you. But yes we've seen stories get changed considerably with no reference to the update. Unfortunately retroactive continuity seems to be pretty much the norm. From a publishers perspective, that's the beauty of the web: unlike print you can change your spin at any time.

It seems that articles on the Guardian[1] have got an "Article history" which list when the article was changed. Saddly they don't tell you how the article was changed.

[1] http://www.guardian.co.uk

Yet another instance where, someday, people will discover and embrace diff. The legal and tax codes are another. Reading through legislation on thomas.gov is like reading through a pile of bad patches with no unified diff context.

Watched your intro video. That seems like a pretty useful service. For the diffs, do you actually convert block elements to text and run the diff(1) algorithm, or do you compare element trees directly? I imagine either way the comparison is fuzzy / heuristic.

We compare element trees directly. It would be easier to extract the text and just do a regular text diff, but this loss of element/tree info means you can't reconstruct html versions of the page with differences highlighted, which for us is an important feature.

Reply on news.ycombinator.com