Tool/method comparison posts are always popular :)

@KamilSJaron You can't ping users unless they've been in this chat room in the past few days (couple of weeks? don't remember exactly) so that user won't see your message.

Instead, leave them a comment under their post telling them to come here.

Oh, and they won't be able to chat if they have >20 reputation SE-wide. If that's the case, ping me and I can give them access (I'm a mod on another SE site which makes ma a chat mod SE-wide)

Ah. Never mind, got it.

@KamilSJaron As expected, not even the dedicated C is faster than grep/fold etc.

@KonradRudolph out of curiosity, why would you say your R approach is the "simplest"? It seems to me it is the most complicated and hard to use answer on the page. Am I missing something?

@terdon Well even discounting the obviously more complex answers in C, C++ and Python … I argue that it’s the simplest because it’s self-documenting due to the use of dedicated library functions

@KonradRudolph Heh. To me it is by far the most obscure and unclear, as well as the hardest to use :)

It may not look simple for somebody who’s not used to the specific API. But somebody who’s equally well-versed in all the languages on the page, this is the most obviously correct solution. All others might do contain flaws

Bah, I can never remember the chat command for strikethrough

foo

Actually using BioPython there’s probably a solution that’s as obvious, if not more so

I don't think anything will ever beat grep -v "^>" file.fa | tr -cd "Nn" | wc -c for speed and simplicity.

Well. For speed, we’ve got the benchmark proving otherwise (although I have some concerns about that benchmark)

For simplicity, I agree for FASTA, but certainly not for FASTQ.

At any rate, I would always advocate using a proper parser instead of an ad-hoc hack (though I violate this rule myself)

@KonradRudolph No, I am only thinking fasta, yes. Fastq is a far more complex program/

Using dedicated tools for fasta is often a waste of time since the format is so so simple.

I would never use a fasta parser for something as simple as that. It would take me more time to implement a script using it than to just run a grep command.

Yes, agreed

@KonradRudolph How so? It's the fastest there.

Ah, no, sorry. The custom C is slightly faster although I really dunno of that's a significant difference there.

The file is too small.

@terdon The C/C++ comparison is definitely off. Heng Li writes good code, but so do the SeqAn folks, and they optimise the hell out of their code

I would be sceptical even if the C code were slightly faster. But a factor 5? That’s just impossible.

I'd need to check the code in more depth. My C-fu is pity-fu so I can't just glance at it and see what it does.

I just can't really imagine how it could do anything better than the super optimized coreutils.

Well, coreutils are optimised for different things. For instance, piping the data is by necessity slower than inspecting it in-place

True.

110 questions before weekend. Does not look very good.

Eh.

Stack Exchange clearly has no interest in this community. There’s very little we can do :/

Dynamics are what they are, ignoring them won’t help

@KonradRudolph Why do you say that? I don't see how they've treated this any different from the other betas.

@terdon They didn’t, that’s the point. They fail to account for the existing competition

As it stands, Bioinformatics.SE is simply not competitive

Oh. Yeah, but that's not really their thing. I mean, SE provides users with tools to launch a new site. If it works, then SE might get involved but we can't expect them to actively advertise all the mini-sites that are attempted. Not before they're ready.

It would be a pity if it didn't go through. I wasn't aware of the public beta until earlier this week, and the peeps at work got really excited, and a few have signed up.

I mean SE didn't choose to create a bioinformatics site, we did.

@Kusalananda Yeah, I still hope but I'm a blind optimist.

> we can't expect them to actively advertise all the mini-sites that are attempted

Sure, but nobody expects that

That said, it would definitely be in their commercial self-interest

… not for every mini-site. Certainly for something like bioinformatics