otabdeveloper

27 ms·

61.
by otabdeveloper 11y ago
> but in practice it should always be combined with a calibration algorithm (which add a trivial O(n) cost to the process). Why not just use logistic regression at this point? The only benefit of Naive Bayes over logistic regression is t
62.
by otabdeveloper 11y ago
No, not quite. _All_ classification algorithms use "reasoning on the average". (A.k.a. "statistics".) Naive Bayes is a poor classifier because it ignores conditional dependency: when having feature A raises the odds, hav
63.
by otabdeveloper 11y ago
systemd depends on Linux-only kernel features. (Like cgroups.) You'd have to patch the FreeBSD kernel first to be Linux-compatible.
64.
by otabdeveloper 11y ago
> I mean, hijacking system logging? What could possibly justify doing that in an init system? DBJ's /service also hijacks logging. In fact, every decent init system hijacks logging -- you want your init to daemonize processes,
65.
by otabdeveloper 11y ago
Every human language is 'tonal', i.e., uses tone for some significant purpose. Some languages also happen to use tone at the phonemic layer of perception. This is a minor difference. Does it surprise you that vowel length in Engli
66.
by otabdeveloper 11y ago
Yes, constexpr in C++.
67.
by otabdeveloper 11y ago
Programmers hate being held accountable, yes. (This is logical given that they control so few of the risks.) But in the end, accountability is the reason why you're being paid that salary, not knowledge of framework APIs or algorithms.
68.
by otabdeveloper 11y ago
Instead of making yet another slightly different Helvetica, maybe somebody should make an effort to make a font that renders Latin, Cyrillic and Greek in a unified style without glaring errors. (As far as I know, there's only one font
69.
by otabdeveloper 11y ago
> what do you really need floating point numbers for? For representing real (i.e., potentially irrational or transcendental) numbers. Irrational and transcendental numbers are actually very common in nature. But you're right that re
70.
by otabdeveloper 11y ago
> he criticizes all of machine learning as simply unprincipled revival of nonparametric curve estimation He's right. 'Machine learning' is just statistics with better, cleaner names for things. The underlying theory is the
71.
by otabdeveloper 11y ago
> terminals are old crusty awful things Not really, they're just poorly documented. They're based on ad-hoc standards that are scattered across many outdated documents. It'd be nice if somebody could actually pull all that
72.
by otabdeveloper 11y ago
I've managed to make it so my memory use is completely deterministic -- no garbage at all.
73.
by otabdeveloper 11y ago
> There are probably other interesting properties you can gain by sacrificing Turing-completeness as well. Not the author, but I wrote another sub-Turing language. In my opinion, there's two useful properties: a) type inference -- y
74.
by otabdeveloper 11y ago
Cool! Another sub-Turing language is my own 'tab', ( https://bitbucket.org/tkatchev/tab ), a text processing language/utility. 'Tab' was born out of a practical need to process text files in a ma
75.
by otabdeveloper 12y ago
If that sort of thing interests you, also check out my language 'tab' ( https://bitbucket.org/tkatchev/tab ). Tab is a statically-typed, functional, type-inferred language that occupies a niche between bash and
76.
by otabdeveloper 12y ago
Where's the SVD? Did I miss something?
77.
by otabdeveloper 12y ago
Use Tab instead: http://bitbucket.org/tkatchev/tab Includes a calculator, but also a complete declarative data-processing language too. E.g.: $ tab "[ log(@),@ : count(5) ]" 0 1 0.693147 2
78.
A compression algorithm designed for code simplicity
1 points by otabdeveloper 12y ago|1 comments
79.
'tab', a new command-line utility for processing tabular text files seriously
10 points by otabdeveloper 12y ago|1 comments
80.
'tab', a new command-line utility for processing tabular text files seriously
2 points by otabdeveloper 12y ago|0 comments