284 karma · joined August 19, 2012
"So you might want to look into not the standard library implementation, but specific locking implentations for your particular needs. Which is admittedly very very annoying indeed. But don't write your own. Find somebody else that wrote one, and spent the decades actually tuning it and making it work.
Because you should never ever think that you're clever enough to write your own locking routines.. Because the likelihood is that you aren't (and by that "you" I very much include myself - we've tweaked all the in-kernel locking over decades, and gone through the simple test-and-set to ticket locks to cacheline-efficient queuing locks, and even people who know what they are doing tend to get it wrong several times).
There's a reason why you can find decades of academic papers on locking. Really. It's hard."
plaud
To applaud.
n. Claim to applause; plaudit; applause.
transitive verb To applaud.For example, if you accidentally put a 'default' on a column when you add it to postgres, it will lock the entire table while it rewrites every row, inserting that default value.
Another common postgres blunder is creating indexes on big tables without using 'concurrently'. This also locks the table for writing and you'll have a bad time.
1. Annual subscriptions is a good idea. Right now we're just seeing if this thing has legs. I didn't consider that people would want to pay a year up front. We are hoping we pegged the average costs per customer accurately. These are well-defined systems with fixed prices, but users are un-predictable. If, on average, users generate way more voicemails than we expected, I don't think we'd have a choice but to reconsider our pricing, or shut down maybe. This ties into your #4 question -- I'd want to make sure everyone could port out their numbers if they want, and collect all their recordings + trancriptions in that event.
2. Right now the number you get is a random US number. This was just a "get it done quick" decision. We'd just need to figure out a good UI for picking your area code, to tack it on. I only hesitate here because if users want a specific area code, in my head that's getting into other product territories (vanity numbers, "local business" numbers, etc), and we're trying to silo ourselves into consumer market (not B2B, there are other products for B2B). We'd also be limited by number availability at the carrier -- maybe it doesn't have the area code you want, kind of thing. Other than that, I don’t think we're against the idea.
3. Right now we store to/from values, duration values, transcription text, and recording files that start when the beep is heard. This data is under lock-and-key in S3 and in a postgres database -- when you play a recording in the UI (or download it someday), you're using a signed URL that has temporary access to just that file. The privacy policy is coming — it’s a good concern to bring up. We didn’t make this to sell people’s data, and want to be up front about that.
4. This is coming, especially for one-off download/export. Depending on the amount of recordings there might be some work on the user's end. It's intensive to zip up a bunch of audio and send it to a user, but not unrealistic. We'd at least have a way for each recording to be downloaded, and a user could have a script run that downloads all their recordings, in a rate-limited fashion.
To put things in perspective, right now there are more preorders than the annual sales of the top 3 or 4 small luxury cars combined (BMW M3, and Mercedes C-Class, etc). I can't recall ever seeing this kind of production capacity needed for a single car model, much less a luxury-class car. Exciting stuff.
1: http://www.bloomberg.com/news/features/2016-03-22/how-tesla-...
[1]: http://www.wired.com/2016/01/gm-electric-car-chevy-bolt-mary...
[2]: https://www.youtube.com/watch?v=g4BIoNsAB8s&feature=youtu.be...
[If] you just want to set up your single project with a task workflow,
then go ahead and use Grunt (or gulp). They will serve you just fine.