brey

16 ms·

121.
by brey 14y ago
if they want to test for domain non-existence, that's what the .invalid TLD is good for. http://en.wikipedia.org/wiki/.invalid
122.
by brey 14y ago
relevant study from the UK: http://www.cyclehelmets.org/1231.html "HEAT suggests that a law making helmets compulsory for cyclists may result in an overall increase in 253 premature deaths – 265 extra deaths from reduced cycling less 12 d
123.
by brey 14y ago
sure, but this doesn't mean the agent model isn't viable for clients with longer-lasting contracts.
124.
by brey 14y ago
right, so an agent needs a lot more clients. but they also have a lot less work to do per client, they're only needed for a few weeks' work every year each. and that serves them well if they build a portfolio of clients with myriad technica
125.
by brey 14y ago
if it's a known malicious entity I don't think I'd want to confirm that my email address is 'live' either way ... the majority use case here is that they're not malicious, merely spammy with good intentions.
126.
by brey 14y ago
in which case you've invested the effort in making the stateful one-time token for unsubscription (is that a word?) - it's EASIER to enact that via a web link than parsing response emails
127.
by brey 14y ago
I believe the data, but the numbers given on that data sheet aren't percentages :-) > Pregnancy rates are the number of pregnancies per 1,000 females age 15 to 19 years 95.46 pregnancies / 1000 females = 9.546% (actually, you can't infe
128.
by brey 14y ago
> where the teen pregnancy rate is around 95.46% (per 1,000 females ages 15 to 19)[1] small point, but this is inconsistent - I suspect you mean 9.546%. a 95% teen pregnancy rate would be somewhat surprising ...
129.
by brey 14y ago
yup. I'd be suspicious of any "unlimited" plan which didn't offer such a guarantee - anything that doesn't put its money where its mouth is is just marketing, a polite euphemism for "lies" in this case. or time-bound it, so everyone's clear
130.
by brey 14y ago
interestingly, that's exactly what Radix sort was invented for - non-comparison sort in O(n) time
131.
by brey 14y ago
nice thing about running crawlers from EC2 - all inbound bandwidth is free
132.
by brey 14y ago
> I used a Bloom filter to keep track of which urls had already been seen and added to the url frontier. This enabled a very fast check of whether or not a new candidate url should be added to the url frontier, with only a low probabili
133.
by brey 14y ago
> Mrs Ebastian's more likely Mr Sebastian ... or was that intentional?
134.
by brey 14y ago
> Yes, 10 years ago it did matter which function to use, because memory and CPU were expensive. Not anymore! entirely wrong. if a candidate doesn't understand the difference between a linked list and a random access array, the differenc
135.
by brey 14y ago
If you get something wrong, and what you type is longer than the expected answer, you end up failing the next couple of questions too. muscle memory - if you type eg "uptime" when the answer should be "free", it fails on the first character
136.
by brey 14y ago
http://en.wikipedia.org/wiki/Betteridges_Law_of_Headlines I'm going with "no" ;-)
137.
by brey 14y ago
are you talking about the barclays PINsentry? if so, it's a chip reader.
138.
by brey 14y ago
> For argument's sake, let's assume there's no such thing as God -- then the entire basis for every religion is fraudulent but it's not -knowingly- fraudulent, which is the major test for culpability.
139.
by brey 14y ago
> try to refrain from saying anything not related to the questions because it can and may be used against you it's a job interview, not a police interview ... show interest in the company, understand completely what they do (the number
140.
by brey 14y ago
these numbers are not all comparable. 1/30 chance of being gay means there's a genetic dice being thrown (as far as we understand). a 1/2500 'chance' of working for google does not have the same degree of randomness. not everyone in the US
141.
by brey 14y ago
you could check the link immediately prior to accepting payment, and show a 'sorry, file is gone :-(' error page if not found.
142.
by brey 14y ago
> ...why not just ask a question about building a real-time monitoring system itself and judge the responses there? agreed. these abstractions for their own sake don't help, make the candidate feel like they're jumping through hoops for
143.
by brey 14y ago
the point of the article is that that detail doesn't matter. short vs long might double the answer, but orders of magnitude are the only interesting answer. however, I feel that the abstraction away from something connected to the job you'r
144.
by brey 14y ago
pricing needs to be much, much clearer, and it needs to be before signing up.
145.
by brey 14y ago
interesting approach. did you consider k-means for this? if so, what influenced your choice?
146.
by brey 14y ago
which conference? I was at the recent Stockholm conference, had a lengthy 1-1 with one of their kernel devs, and I picked up the opposite impression.
147.
by brey 15y ago
I buy a lot more pens than suits. it's worth optimising cost on regular purchases far more than on very infrequent payments.
148.
by brey 15y ago
oh, this is good. just what I've been looking for.
149.
by brey 15y ago
> Hackers, and thus by definition the people with the most reasonable claim to define the term, having done so originally. and in the first half of the 20th century the word 'gay' meant 'happy'. shall we insist that's still the proper m
150.
by brey 15y ago
or, as has been used in scientific papers since forever, an abstract.

More ›