Ask HN: An ideea about a link checker

11 points by sirrocco ↗ HN
I've been thinking lately about building an application takes in an url and basically crawls that entire site.

It would be an easy way of finding problems with the a site (like 404, 500 ...etc).

I tried the webmaster tool from google - but I know I have a 404 link on my site , yet it didn't find it. I also tried some other sites that do crawl a website but they had a limit of 2-300 links and then it would stop. You would then have to buy an app that was a bit expensive.

I'm thinking a web app where you can point it at your site and just receive a report when it's done.

Would anyone need something like this ?

15 comments

[ 4.2 ms ] story [ 20.5 ms ] thread
this is on the list of things I have wanted to do at some point.

I believe a wget -spider should help you find any 404's, but I wanted to have each link validated as well, and itd be nice to have as a simple web service

Yea, it could validate the html on the page, could show some statistics - like linkcount, time it took to download the html ( it could then check for the links of the img tags )

I was thinking that there could be a plan wher you would have a scan a month to see if any problems appeared in the meantime.

Offer varying frequencies and opt-in tests. This becomes a "freemium" model very easily.

Here's a challenge: Could a service also try to detect common problems with appearance and rendering? Can this be done without some kind of AI akin to OCR? Certain things would be easier to detect, like content text being overlapped by another element.

Maybe just Mechanical Turk it?

I do get the added-value of having it as a webapp for some but as a developer, I'd rather use Tarantula in my test suite. Hunts down 404s, 500s but can also do HTML validation, check against common attacks (CSRF, XSS, etc.)

http://github.com/relevance/tarantula

nice, didn't know about that one.
Then you probably also want to hear about SpiderTest. http://github.com/courtenay/spider_test

It's more of a strict crawler but also runs prototype's Ajax.updater javascript calls — which can help hunt down AJAX-only actions that cause 404s or 500s.

Nice, thanks for sharing. Do you any similar Python software?
Nope, not out of the top of my head. Though designed for Rails or Rack-based platforms, Tarantula doesn't seem to be too heavily linked to it. I'm guessing with some minor Ruby work, you can probably run it against any website.

It just emulates a browser session and doesn't actually interact with the server-side code, just your front-end HTML.

Do your research well. There are quite a number of such apps out there already.
It would be great if you could give a couple of links, I did search but didn't find something that made me think : ok - they are doing this so great that it's next to pointless to even start.

The ones that I found , I didn't really like - which is I'm even asking here - If I can't find others doing this , maybe nobody wants something like this.

Something with a spell checker would be nice.

Edit: actually, no pun intended, even though there is a spelling error in the title.

Yeah, my bad about the title. But I don't really see a spell checker as something you want . If you have .. 1000 pages and I find 1 error in 10% of them, it would be a lot of work to correct that work. Not sure anyone would actually go and correct them.

But it could be a premium service i guess.

(comment deleted)
Hi everybody. I'm from Linkvive and we've actually been building such a service for the last few months. We're very excited to see interest in it on HN and look forward to sharing our service with you guys soon. Signup at our form (http://linkvive.com) and we'll send you a single e-mail when it's live in the next few months. Cheers!