« first day (321 days earlier)      last day (1214 days later) » 

02:48
@ngn are you interested in kdb job?
03:00
$[interested; (""sv inverse myname),"@gmail.com"]
$[interested; (""sv inverse myname),"[email protected]"]
 
4 hours later…
ngn
ngn
07:16
@HaoDeng of course, i'm interested, especially if remote work is ok and you're not in a hurry - i'll be on a job hunt later this year (not right now). i'll send you an email today.
07:26
There is a k9?
@rcabaco the version available at shakti.com, which is different from the release on anaconda is called like that by some unofficially.
@ktye Oh, i didn't know. So the new releases are now at shakti.com instead of anaconda?
@rcabaco it seems so. but they are quite new and not as mature just yet
@ktye you mean it's a new implementation?
@ktye (sorry for all the questions)
@rcabaco yes, and some adverbs have different meanings. that's why it needs a new name if you want to discuss k code.
ngn
ngn
07:42
the obvious question is: why no k8?
And i can't help but think that moves like this push aside whomever was interested in using the language. It has nothing to do with open source or any of that, even the closed-source languages make an effort to keep the community in the loop in some way. But then again, this is my problem, not theirs.
@ngn this is one of these mysteries. i heard it first from john: news.ycombinator.com/item?id=21925761
@ngn to finish off the raw input conversation of the other day, in q: code.kx.com/q/ref/dotz/#zpi-input
@ngn but it still is line processing (needing a return)
@ngn Dyalog APL has ⎕ARBIN which lets one read directly from a file descriptor a number of bytes
ngn
ngn
07:57
@rcabaco i thought they had deprecated that.. apparently not
@rcabaco in q, i would have expected read1 with a pair of file descriptor and byte count, or equivalently in k4: 1:(fd;nbytes)
@ngn I couldn't understand how to make ⎕ARBIN work for stdin. I think it might be used for ⎕SR. Probably they are still there for legacy reasons.
@ngn Didn't see read1
ngn
ngn
@rcabaco that's just monadic 1:
@ngn Yes
@ngn maybe helps avoid Stack Overflow conflict with kubernetes
@chrispsn i don't think anyone cares about name conflicts, who spends a single character for a name.
ngn
ngn
08:12
k9=canine :)
Maybe ngn/k is k8 ;)
ngn
ngn
reminds me of a company i worked for long ago, they went from 4.6 to 6.0 for marketing reasons
the competition were 1 version ahead and the stupid customers thought these numbers must mean something
or maybe the marketing team was trying to justify its own existence with reasoning like this.. i'll never know :)
@ngn you are not referring to netscape, are you?
ngn
ngn
@ktye it's not netscape
wikipedia: In November, work on Netscape 5.0 was canceled in favor of developing a completely new program from scratch.
so N5 existed, but got rewritten - that sounds like a valid reason to renumber it
@rcabaco i don't like the design of 1:, to be honest. in some cases the n in 1:(f;n) is offset, in others it's length
(where f is a file-like thing and n is an int)
09:09
@ngn "read1[x], where x is: an file symbol or a list (file;offset;length) or a list (file;offset)"
ngn
ngn
@rcabaco ..and later: "a list (fifo;length), returns length bytes read from fifo"
@ngn oh, true
with named pipes
ngn
ngn
@rcabaco did you know that k4 uses select(), not read() to read from stdin?
@ngn no
@ngn implications?
ngn
ngn
@rcabaco you can't read() from it, i guess (i'm not very familiar with select())
ngn
ngn
09:38
@rcabaco if you can find geocar somehow.. i know he knows this kind of stuff. he's been active on hn lately.
09:57
@ngn both read1 and read0 work but neither reads only 1 byte. they read input till newline and cut to the specified length.
10:10
@ngn his email is in the Shakti google group
 
2 hours later…
ngn
ngn
11:44
@user689349 hi! if you want write access, please see this message

« first day (321 days earlier)      last day (1214 days later) »