/dev/chat

« first day (1392 days earlier) ← previous day next day → last day (3555 days later) »

Seth

12:01 AM

No, I meant which answer are you saying is correct?

Braiam

7 mins ago, by Seth

I don't know how the swap file answer got so many upvotes..

that and noleti's

I was about to write a single answer with both contents, figuratively

Seth

yeah, he's right.

I misunderstood what he was saying in the second half though. Makes sense now.

11 hours later…

Pandya

11:16 AM

What is relation between FSF(Free Software Foundation) and GNU ?

What is kernel and it's importance/relation with operating system?

Baard Kopperud

@Pandya FSF is the organization started by Richard Stallman. GNU is a project started by Richard Stallman to create a free Unix-like operating system. Though I guess FSF (as an organization) do/organize much of the work on GNU too.

@Pandya The kernel is the most important part of the operating system - the part that is actually forms a layer between the computer hardware and the software you want to run on it. But you need more than just the kernel to get an operating system... At least you need some programs to communicate with the user (a shell like bash or some graphical user interface), and some programs to do simple things (like copying, deleting and renaming files, and editing files).

4 hours later…

Faheem Mitha

3:12 PM

@Pandya This information is easy to find on the net. Start with the FSF web site. Wikipedia is also a good source for such information.

Should I give yet another answer to this question?

This answer could be improved, but has the basic fact right, which is that lintian is a tool whose main purpose it to check packages for compliance with Debian Policy. The people who work on Policy and lintian are largely the same. One point that is not clear is that lintian is a command line tool which can be run either during the building of a Debian binary package, or on an already existing Debian binary package, and will return the results of its tests to standard output. Lintian is automatically run during builds of the Debian archive, and its results can be found on lintian.debian.org. — Faheem Mitha 9 mins ago

I added a comment on one of the existing answers, in the hope that the answerer would add it to his answer, but he does not seem inclined...

But there are 3 answers here already. Should I add a fourth?

6 hours later…

Gilles

8:51 PM

Interesting: data.stackexchange.com/unix/query/213319/… data.stackexchange.com/askubuntu/query/213319/… data.stackexchange.com/superuser/query/213319/…

Questions on U&L get more views than on AU or SU during the first few months, but after that the trend goes the other way

I find this intuitive, but it's nice to validate my intuition with data

U&L has fewer questions and pays more attention to each individual question when it comes out

but in the long term the sites with the most amalgamated views get cited more and so get more views overall

U&L and AU start out with the same standard deviation but later AU has more deviation (more differenciation between oft-cited questions and forgotten questions)

SU has more stdev from the start, that's consistent with its having a broader topic

it would be interesting to see the numbers without the high-scoring outliers but that's beyond both my sql and my stats

Braiam

9:07 PM

@Gilles why I'm not surprised?

Gilles

9:55 PM

Do we have a tag that's like parallelism (which I'm creating), but meaning distributed across machines rather than on multiple cores on a single machine?

distribution is already taken for something completely different

Super User uses distributed-computing

Server Fault uses distributed

Braiam

10:25 PM

@Gilles really?

casey

cluster or something involving that word? distributed-computing gives me the idea of de-centralized computing (seri@home, etc) whereas multi-node parallel computing is done on one big multi-node machine. Also an MPI program that would run on such a computer really doesn't care if it is running 8 instances on one node or 64 across 8 nodes, so I'm not sure a distinction is really needed

Gilles

@casey as you say, cluster is more restricted than distributed

casey

e.g. I run my code the same way locally on 12 cores as I do on yellowstone on my 512 core allotment

@Gilles true, but I suppose it depends on the type of questions that tag will be meant to intercept. If I were going to post a question about getting MPI setup on a few machines, or perhaps setting up PBS, SLURM or HYDRA to spool those jobs, I'd first look for variations of parallel, mpi, cluster before thinking of distributed

but I will plainly admit I'm biased toward that application of multi-node computing than the wider distributed de-centralized computing

Gilles

10:55 PM

What is it today with people asking “I have some input and I want to produce some output, how do I do it?” and refusing to us what the input and the output look like or how they're related?

and with people posting obviously-unhelpful or dangerously misleading answers, and getting upvotes

Braiam

@Gilles give up fixing humanity

Gilles