I think they might be suffering heavy load; I don't think waiting an hour is an acceptable delay for web dev - it would make for a very slow iteration loop!

  > There are 68 other webpages queued before this request. 
  > Your request should be completed in about 1 hour.

We are suffering an overload right now. Normally a test does run in 20-30 seconds. We're doing everything we can to serve all the users that have signed up today and scale our servers.

Sincerely,

Browserbite team

I have results - I'm impressed with the quality of the output - especially that errors aren't cascaded! I would spend more time looking at it now but unfortunately it takes a while to load each screenshot currently. I'll check back tomorrow when your servers have had time to recover.

Do you have plans to support different (user configurable) window sizes? My site uses media queries to differ the appearance based on browser size, which means the older IEs can't help but fail to look the same as modern browsers.

Great work, and good luck scaling it!

We have plans to support the typical resolutions as well. 1366, 1920 and typical mobile/tablet resolutions. Is that what you're looking for?

P.S. You can use the feedback button in the app to send more idea!

I was expecting there to be an option to upgrade to a paid account and be able to skip the queue, but it doesn't look like that's the case.

I don't know if they were predicting a HN size load in this stage of their development. When I interviewed Kaspar for this article he said that they're looking for funding this fall to help scale, which looks like a good idea. I found the service interesting though, well worth a bookmark!

We ARE working on the payment module right now to give paying customers priority. For now, we just try keep our servers up to handle the load :)

Kaspar from Browserbite

Agreed, I would pay for priority queue in a heartbeat!

Kaspar, can you talk more about how a vision algorithm is used? It sounds like it's more involved than stitching together screenshots of the whole page and diffing them.

You can write a whole PhD thesis about it. Basically it is a result of a 2+ year research project how to detect differences that are relevant to people. In fact the computers find a lot more differences when you compare pixel-by-pixel. The keywords are machine vision, artificial intelligence, machine learning.

The hardest part of working out the right balance of what to detect was to make the algorithm less picky without losing an eye for detail.

We've benchmarked the algorithm against professional layout testers and the fault detection quality is on the same level with them.

Will I be able to run the diff against previous versions of the site? This is a huge pain point for us and I've been looking for a service that will let us automate testing for CSS breakage on deploys.

Yes, it is possible! Can you drop me an e-mail (you can find it on our site) on how do you plan to use the service for this purpose. Skype call could be best!

Thanks, I'll follow up when you're not so busy.

Worked pretty well for us, already spotted a couple of funny css bugs on our home page on IE7 (of course)

Kaspar, do you have a general idea of what pricing will be when you start offering paid tests?

We're planning to have monthly subscription that would include 50-100 comparisons (depending on how many configurations you want to have in the package) for $99 per month. Priority processing is for all paying customers, of course.

There's also going to be Enterprise pricing that includes - scripting support, testing behind firewalls and some other extras.