tea.mathoverflow.net - Discussion Feed (Learning Roadmap for Software Engineer)

Jacques Carette comments on "Learning Roadmap for Software Engineer" (6352)

2010-06-23T19:04:53-07:00

@Steve: that's a fair answer. I definitely agree that the language on MO should not change. And if people agree with Andy that theoretical computer science is indeed 'solid mainstream mathematics', then indeed the topic should be of interest to 'research mathematicians' (even if some typically label themselves otherwise).

Steve Huntsman comments on "Learning Roadmap for Software Engineer" (6351)

2010-06-23T18:36:43-07:00

@Jacques--As I mentioned in the thread you just started, I think TCS is absolutely fine here, and I personally enjoy many of the questions on algorithms (if not the more abstract stuff like grammars or what not). I actually own (almost all of) a network security company that has made use of some nontrivial mathematics at various points, so I am very sympathetic to the CS and especially TCS folks. But I don't think that (beyond admitting what I think is the obvious fact that TCS is a branch of mathematics) there should be any formalized attempt to cater to folks that aren't willing or able to speak the language here. We are providing a service already.

Andy Putman comments on "Learning Roadmap for Software Engineer" (6346)

2010-06-23T18:14:10-07:00

Theoretical computer science is solid mainstream mathematics, and I hope that researchers in it feel welcome here! However, it is a very different subject than software engineering.

Jacques Carette comments on "Learning Roadmap for Software Engineer" (6345)

2010-06-23T17:36:50-07:00

@François - thanks, I had not seen that TCS proposal. I am definitely a follower now! [Not that I want to 'leave' MO, but this is still a good idea to have as well]

@Steve - TCS is as much about code as astronomy is about telescopes, or math is about natural number arithmetic. TCS is the one area where 'applied math' does NOT mean DEs! It does mean algebra, combinatorics, graph theory, category theory (even the n-cat kind), number theory, and so on.

François G. Dorais comments on "Learning Roadmap for Software Engineer" (6341)

2010-06-23T17:02:39-07:00

Jacques, you might want to check out this TCS proposal on Area 51. That said, I've been happy with the TCS presence on MO; I think it does have a place here.

Steve Huntsman comments on "Learning Roadmap for Software Engineer" (6339)

2010-06-23T16:57:16-07:00

@Jacques--There are many other consumers of advanced mathematics. Chemical engineers may deal with quite complex ODEs and PDEs; civil engineers may deal with finite element or lattice models of mechanical deformation and fracture; electrical engineers may deal with filtering/state estimation, information and coding theory, sparse linear algebra, probability, etc.; some bioinformaticists deal with highly nontrivial combinatorial constructions, and I haven't even mentioned physicists, computer scientists, economists, or for that matter autodidacts. It is highly unlikely that a single site evoking MO could possibly get the incentives right for all of these parties and the mathematicians. If it was only computer scientists there might be a quid pro quo whereby mathematicians could ask for and get useful bits of code, but that won't really help much given the other disciplines that are likely to factor in. What would a mathematician possibly want a chemical engineer to do in exchange for providing background on the BZ equation? Instead it will almost surely be necessary for others to learn to speak the language (enough mathematics to ensure a favorable response to questions) and to offer the coin of the realm (questions that are sufficiently nontrivial that they will be worth the time invested in answering them).

Jacques Carette comments on "Learning Roadmap for Software Engineer" (6336)

2010-06-23T16:36:42-07:00

I think one of the reasons you get such questions is that MO is the only place on the net where people who know serious math hang out. In particular, there is absolutely nowhere else that someone who is really heavy-consumer of mathematics (like a lot of 'modern' computer scientists and the formal-methods variant of a software engineer) can get decent answers. I know a number of my questions could not have been answered anywhere else, because they lie in the nebulous region between math and computer science.

Basically what is desperately needed is a MathOverflow for theoretically-minded computer scientists. I was hoping that MathOverflow could actually fulfill that role itself, since "theoretical computer science" is (to me!) just a slightly different flavour of mathematics.

supercooldave comments on "Learning Roadmap for Software Engineer" (6242)

2010-06-22T00:55:48-07:00

I just wanted to add my two cents worth, being the one who provided the rather large answer to this question. I answered because the question fell within my field of expertise (theoretical computer science) and I could, off the top of my head, provide a decent answer to the person's question. On the other hand, the question was vague and undirected, so my answer may have been off target. Furthermore, the original author has not taken the time to respond to the answers provided or even to fix the original question (but it has only been 24 hours since the original posting).

In general, however, I agree with the sentiment that there are too many of this sort of question.

Andrew Stacey comments on "Learning Roadmap for Software Engineer" (6240)

2010-06-22T00:22:52-07:00

My gut feeling is that there have been a few too many of this type of question recently. My interpretation of the "of interest to research mathematicians" is "questions that research mathematicians would like to ask each other about actual mathematics". Occasionally, we get questions that fit that sentence but with the alternative meaning of "about" and those I tolerate so long as there aren't too many of them.

Also, I don't know anywhere else on the internet that would be comparably good for asking such questions.

The reason for that is simple: anywhere where it's the norm to ask such questions is somewhere that research mathematicians aren't going to spend much time on! If someone not in the main academic stream is really interested in learning some mathematics "on the side" then they really ought to do the following: find a nearby university with a mathematics department, go there and ask to make an appointment with a mathematician - offer to pay for their time! - and ask them directly what they recommend. The mathematician will be able to give a much more focussed answer and so much more useful. Plus an initial contact will have been made which, if not abused, could be much more useful to the questioner. I would wonder at someone's understanding of what real mathematics is if they think that an answer to a general question on MO is really going to help them learn mathematics.

It's alright for someone who is already in the system to ask such a question, because all they really want is a "where do I start" - they have the support and network to continue. Of course, I would hope that they had asked in their own department first and I would hope that they had gotten the response, "I don't actually know about that area, but if you can find a place to start, then I'll be happy to help where I can.".

Akhil Mathew comments on "Learning Roadmap for Software Engineer" (6238)

2010-06-21T18:28:13-07:00

I don't personally have a problem with occasional such questions (but, then again, I must state my bias: I have asked one such roadmap question (albeit for a specific subject) before!), even though they are not the purpose of MO. A lot of users of MO are students (or have students), and the responses are likely to help many beyond the original posters. Also, I don't know anywhere else on the internet that would be comparably good for asking such questions. For instance, you can't get advice from research mathematicians on Aops in the same way. Surely there are people who'd like to become serious mathematicians, and are thinking about graduate school, but might (for instance) be at a liberal arts university without a graduate math program, and no easy way to ask active researchers.

However, it seems that they become redundant fast.

Gjergji comments on "Learning Roadmap for Software Engineer" (6237)

2010-06-21T18:19:17-07:00

I believe we have had questions "a learning roadmap for subject X" before, and I think these are more appropriate than "a learning roadmap for person Y". I didn't vote to close but I feel like the second question incorporates the first one, i.e. an answer should outline a few subjects appropriate for person Y and then give a list of references for each subject. A question of the first type is, on the other hand, just asking for advice from a specialist in a specific subject, which is what many users expect from MO anyways.

Andy Putman comments on "Learning Roadmap for Software Engineer" (6236)

2010-06-21T16:53:02-07:00

While I can't say that I'm thrilled with questions like this, I'm not opposed to their existence (as long as they don't start showing up too often).

I have definitely learned things from answers to questions like this that help in advising students. That strikes me as a reasonable justification for their existence on MO.

Grétar Amazeen comments on "Learning Roadmap for Software Engineer" (6235)

2010-06-21T16:49:34-07:00

In my opinion this is not a question for MO. I cannot see how this is a question of interest to research mathematicians.

Qiaochu Yuan comments on "Learning Roadmap for Software Engineer" (6234)

2010-06-21T16:41:36-07:00

http://mathoverflow.net/questions/28916/learning-roadmap-for-software-engineer has garnered two votes to close and at least one upvote, but nobody's yet explained their reasons for closing in the comments. I'm not sure what to think about these kinds of questions. We've had similar ones in the past and they've always struck me as a little broad and a little off-topic; this is precisely the kind of question a research mathematician wouldn't ask! On the other hand, people are answering, and the question was asked in good faith. What's the community consensus?