3 ms·

>don't parse HTML with a regular expressions Even if I agree with you (I don't know because I've never been in the situation where that was a thing I was consi

>don't parse HTML with a regular expressions Even if I agree with you (I don't know because I've never been in the situation where that was a thing I was considering and weighing against other options) what I'm arguing here is that if someone asks how to do that then you should either tell them along with the admonishment or not say anything at all. But don't go around being a dick by yelling "haha that's the dumbest thing ever and I'm not going to tell you because of best practices"

Pamar 8y ago
While I agree with you in general (i.e. educate people or just stay silent instead of pontificating/be condescending if not rude), what would you say about the canonical answer on StackOverflow, then? https://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags/1732454#1732454 https://stackoverflow.com/questions/1732348/regex-match-open...
- mlevental 8y ago
  meh. the tone isn't so bad but, while at first i thought otherwise, it doesn't really comprehensively explain why it's a bad idea (on first skim i thought most of the content of the answer was an explanation of the difference between a state machine and cfg or something like that). i think a good model for how to answer questions like this is really the same as you'd answer kids' questions as a parent (or maybe anyone's questions?): a flat out no is completely unsatisfactory and lazy. a no with a measured explanation is good.
  - Pamar 8y ago
    Again, agreed. But what when something is at the same time pretty much wrong but also something that everyone tries to do because superficially looks like a good idea? The reaction on StackOverflow was because they were getting dozens (hundreds) more or less identical copies of the same question. It was the same with "parse email via regexp" but I understand that more recent versions of the RFC make this doable.

Pamar 8y ago

While I agree with you in general (i.e. educate people or just stay silent instead of pontificating/be condescending if not rude), what would you say about the canonical answer on StackOverflow, then? https://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags/1732454#1732454 https://stackoverflow.com/questions/1732348/regex-match-open...

mlevental 8y ago
meh. the tone isn't so bad but, while at first i thought otherwise, it doesn't really comprehensively explain why it's a bad idea (on first skim i thought most of the content of the answer was an explanation of the difference between a state machine and cfg or something like that). i think a good model for how to answer questions like this is really the same as you'd answer kids' questions as a parent (or maybe anyone's questions?): a flat out no is completely unsatisfactory and lazy. a no with a measured explanation is good.
- Pamar 8y ago
  Again, agreed. But what when something is at the same time pretty much wrong but also something that everyone tries to do because superficially looks like a good idea? The reaction on StackOverflow was because they were getting dozens (hundreds) more or less identical copies of the same question. It was the same with "parse email via regexp" but I understand that more recent versions of the RFC make this doable.