asm

Clarification on the title of this post; Phan is written in PHP7, but will analyze code written in any version of PHP.

Analysis takes a lot of CPU and we're not doing any parallel processing. Phan itself is about 30k lines of code and takes about 1.5s for a clean run.

That being said, you should take a look at the '-s' flag for saving state to a sqlite database. After the initial run the stored state can be used to only scan changed files. For a patch with a few files changed it should be about 0.5s to run the analysis.

I'm not sure. PHP7 is pretty new so package maintainers may not have had a chance to put binary releases together. You may need to compile it yourself from https://github.com/nikic/php-ast.

Definitely.

If you want to start working on that, feel free to open up an issue (https://github.com/etsy/phan/issues/new) to start a conversation about it.

We currently store method/function callers in the sqlite database (see https://github.com/etsy/phan/blob/master/src/Phan/Analyze/Br...), but aren't yet exposing an interface for getting callers. We'll want to store callers for constants, classes, properties and global variables as well.

Phan does not attempt to make sense of things like `$class_name->$function_name()` or `$$v`. I'd be happy to see a pull request that could handle simple strings, but otherwise I imagine the best we can do is emit a warning.

I'm not sure what SIMD refers to. Can you clarify that?

We'd like to get there eventually. Phan populates a sqlite database when analyzing code (if run with the -s flag). This allows fast runs on single file changes or patches. We're hoping to use this for things like auto-complete and finding callers for a class/method/whatever eventually.

For now, we have a primordial Syntastic vim plugin (https://github.com/etsy/phan/blob/master/plugins/vim/bundle/...) for getting errors into the quickfix view.

Co-author here. Happy to answer any questions folks have about Phan.

We used construction paper to black out the back around the tablet. My guess is that the paper wouldn't like the humidity but the rest would be fine. Perhaps you could swap in some black plastic material that would work with spray adhesive.

Hi, I helped hannahmitt make the mirror.

The rectangle of light isn't noticeable at all in normal light if you adjust the brightness on the tablet well. In a dark room it can be noticeable, though.

Unfortunately, the touch screen doesn't work through the mirror, however, we're using reusable sticky stuff to attach the mirror to the tablet, so its easy to remove if needed.

It was written to scratch a personal itch; never as a business. I'd be happy to see someone else pick it up and do something with it.

http://snapnote.io/

fixed

That you're asking about keeping it unguessable makes me think that it feels like too few. Its actually only 5 bytes of random (the first character is always zero). 2^40 is a pretty large number and I don't expect to get high coverage over that space any time soon. That being said, perception is important, so I should probably up it.

As for hot-linking; I have no business plan and just wanted to make something useful. I should probably figure that stuff out.

Mobile devices were very much an afterthought. I was getting mobile traffic and did a best effort to make it passable, but I don't recommend it.

You should check out Bugshot (https://itunes.apple.com/us/app/bugshot/id669858907?mt=8) which I think does a much better job on iOS.

That's a good idea, I'll do that. Thanks.

I totally agree. Paste is currently supported on Chrome and its been super useful. I'll see if I can get it supported elsewhere. Thanks.

The point is that we're not looking for specialists in any particular language.

To cure the curiosity, however, we're largely running on Linux, Apache, MySQL, PHP with many other tools thrown in the mix.

re Quixotic Behavior: We want people who are comfortable taking a stand on an issue and going off on a mission to solve the problem. Not every great idea seems reasonable before people see it working.

re Full-Stack: We're looking for people who design and build full systems from low to high levels. Some have made their career working just as a front- or back-end hacker. We want to meet people who wouldn't dream of letting someone else take half their work or who would be comfortable throwing part of the problem over the fence. People that are a good fit probably don't care that much about what the stack is beyond some reasonable constraints.

If you end up sending that cover letter there's a chance that we'd all be amused enough to read your resume.

At Etsy, our coffee pot is hooked in to our IRC bot so that we can ask how many cups of coffee are left ("?coffee"). Someone from the ops team just opened a bug report due to non-compliance with RFC 2324.

The IRC coffee service uses a Pelouze postage scale and https://github.com/morria/coffeed

In the beginning of the article, there seems to be an implicit claim that this kind of testing is bad because those who score on the border of a "cutoff" only have a probabilistic chance of making it into these coveted schools. The author seems to be suggesting that the error prone nature of measurement instruments means that we need to create something that is free of error or we need to stop using them all together. Unfortunately such instruments will never be error free and there doesn't seem to be an acceptable alternative for collecting predictive information. The measurement community will continue to improve on their instruments, but it seems like the only real remedy given the author's criticism is for these schools to just admit everyone so that nobody misses out. This, of course, isn't realistic.

When resources are thin and there is no way to deduce a provably optimal allocation, we start having to make guesses given the best possible information and algorithms. Until we can come up with something better, schools will continue needing instruments like these (as well as other tools) for making the best possible guesses.

Children's Progress, a small New York City based data-driven-education company, is hiring.

http://www.childrensprogress.com/about-us/softwareengineer.s...

Perhaps an assessment of learning ability rather than achievement would distinguish students who could get the most out of a scholarship.

Perhaps by assessing all students to find the easiest concept the child has not mastered, teaching it to them, and then measuring their ability after being taught, we would be able to measure their ability to use new information effectively.