Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
samuell
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
54 ms
·
631.
▲
by
samuell
9y ago
I made a udemy-course called "Drastically improve your speed on the Linux commandline" [1]. It has generated enough to cover the a bunch of domain names, but not much more, and I have many times questioned whether it was worth bur
632.
▲
Pony: Combining safe memory sharing with Erlang-like actors
(tutorial.ponylang.org)
202 points
by
samuell
9y ago
|
62 comments
633.
▲
by
samuell
9y ago
I see there's a playground [1], and a pres video from this year's curry on conf [2] available. [1] http://playground.ponylang.org/ [2] https://www.youtube.com/watch?v=HGDSnOZaU7Y EDIT: Oh, and got
634.
▲
Pony Programming Language
(ponylang.org)
5 points
by
samuell
9y ago
|
1 comments
635.
▲
Ask HN: Problems needing to be solved in bioinformatics? (Aka startup ideas)
1 points
by
samuell
9y ago
|
0 comments
636.
▲
by
samuell
9y ago
This. I tend to think a key to really understanding something, is the ability to put that "something" in to contexts at a wide variety of abstraction levels. That's what I hated the most with uni: Some professors could write
637.
▲
by
samuell
9y ago
For a recent book / set of papers on the topic of understanding biological information, I can recommend having a look at the BINP book: http://www.binp.org (Biological Information, New Perspectives). It is a collection of p
638.
▲
by
samuell
9y ago
... and then trust that this designer- and intention-less stochastic, random, evolution brought us brains that tell us the ultimate truth? Given the heavy biases of the anthropic principle, odds don't look good (... when we can't
639.
▲
by
samuell
9y ago
I maintain that it takes utter foolishness to not acknowledge the screaming intelligent design behind nature. Seldom does it get clearer than when we as information systems designers study the information systems in nature, and find basical
640.
▲
by
samuell
9y ago
Perhaps they include the weight of stuff like histones [1], which DNA is wound up upon in order not to get entangled. That could add some considerable weight. Also there are various epigenetic modifications on top of that, further adding we
641.
▲
by
samuell
9y ago
Well ... it seems you might be right. Another estimation states: "Capable of storing 215 petabytes (215 million gigabytes) in a single gram of DNA, the system could, in principle, store every bit of datum ever recorded by humans in a c
642.
▲
by
samuell
9y ago
Yep, the smallest atom, hydrogen, weights ~ 6 * 10^-23 of a gram [1] (that's 0.000... some 23 zeroes ... and corresponds to 1 "atomic unit"), and other atoms weight just tens or hundreds times more (so still in same ballpark)
643.
▲
by
samuell
9y ago
Yes. I've been hopping around on too many distros in the past and always come back to LTS versions of xubuntu. Everything else just creates a lot of unanticipated pain or annoyance after some usage. Lubuntu would be the next best, if y
644.
▲
by
samuell
9y ago
From the crowd-sourced intr feature table I initiated [1] it turns out light-weight threads is one of the the more prominent features that is only supported by Crystal, and not Nim. [1] https://docs.google.com/spreadsheets&#
645.
▲
by
samuell
9y ago
Thought the same. The first "screen" that's big enough for reasonably fit a playable piano keyboard.
646.
▲
by
samuell
9y ago
I researched basically all the solutions (wikis, tiddlywiki, evernote etc etc), but they all had some limitation either in offline access, shareability, portability, openness or something else. What finally made it into a working solution i
647.
▲
by
samuell
9y ago
And then the trio sonatas for organ are the master pieces of the master pieces if you ask me ... with three truly independent melodies masterfully interwoven most of the time ... did only find one so far though: http://allofbach.
648.
▲
by
samuell
9y ago
If any Bach newbies want something easy listened to start with, don't miss the classic, Toccata & Fuga in D minor: http://allofbach.com/en/bwv/bwv-565/ Especially the middle parts (2:37 and forward)
649.
▲
by
samuell
9y ago
The explanation of FPGAs between 5:00 and ~12:00 was very very good. TL;DW: - FPGAs are more power-efficient than CPUs and GPUs - FPGAs are more flexible than ASICs FPGAs vs CPUs: - CPU: Sequentially apply instructions on data in place - FP
650.
▲
by
samuell
9y ago
Tried to scan the video to see if they have any support for reconfigure.io [1] (Code FPGAs with Go concurrency primitives). Didn't see any mention. Anybody knows more? Or I'll have to watch the vid in detail ... looks pretty pedag
651.
▲
by
samuell
9y ago
Suggestion: Keep an eye on Go as an emerging data science language (See http://gopherdata.io/ for more info)
652.
▲
by
samuell
9y ago
Philips CDE tool [1] looks like such a beautiful thing! > "Prepend any set of Linux commands with the "cde" binary, and CDE will run them and automatically package up all files (e.g., executables, libraries, plug-ins, conf
653.
▲
by
samuell
9y ago
SciLuigi adds a simple level of audit logging to Luigi [1] (See the AuditTrailHelpers mixin in https://github.com/pharmbio/sciluigi/blob/master/sciluigi/au... for details), together with an API that
654.
▲
by
samuell
10y ago
For anyone interested in learning more, or helping collect the best Ada links, I started out on an "Awesome Ada" list some time ago: https://github.com/samuell/awesome-ada
655.
▲
by
samuell
10y ago
Ada ticks a fair amount of the boxes in the "Interesting statically compiled languages by wishlist features" google doc, which I started, and had some crowd-sourcing help to fill in: https://docs.google.com/spreads
656.
▲
by
samuell
10y ago
Yep ... only .com instead of .co.uk Strange that bbc doesn't have some resolving setup for that ... dupes supposedly being bad for SEO and all ...
657.
▲
by
samuell
10y ago
If more developers had a second look at FreePascal/Lazarus [1], they'd realize they could do pretty much the same thing, but with native compiled speed, minimal memory usage (no bundled JREs) on all the desktop platforms. It'
658.
▲
Program AWS FPGAs Using Go
(reconfigure.io)
1 points
by
samuell
10y ago
|
0 comments
659.
▲
by
samuell
10y ago
We found great success with the LIBLINEAR SVM implementation [1] though: Extremely good performance, to the point that it affects scalability too, with predictive performance acceptably close to libSVM with the RBF kernel, for a large che
660.
▲
Gobot 1.3
(gobot.io)
3 points
by
samuell
10y ago
|
0 comments
More ›