Among many other things the ACM sponsors the International Symposium on Computer Architecture (ISCA) which is occassionaly derisively called the "International Symposium on Cache Architecture."
@Raphael Huh. Works for me. It's just a pointer to the ACM computing classification system. (The first major subject is "hardware" and includes caching and pipelining.)
@Raphael The question is literally "How is the L1 cache implemented to support pipelining?"
The Intel i7 and ARM A9 are used as particular references to support the claim that anyone ever talks about pipelining a cache.
@Raphael Sure. It's not a particularly well worded (or well researched) question, but that doesn't make it more appropriate to dump on ee.se.
I'm perhaps feeling a bit sensitive today.
I think it is extremely likely that the os.se beta will be stopped today due to inactivity.
(Which is not surprising.)
But I think the fact that os.se got to beta demonstrates that cs.se needs to be more receptive to questions that are computer systems related.
Systems is, perhaps, more clearly an engineering discipline than is theory.
And one of the things that academic researchers in systems often publish papers about involves comparing specific artifacts.
If we can't talk about specific implementations then we can't talk about what people discovered during the process of implementation. (Which may confirm or refute some theory they had about system organization.)
And all we are left with is questions from a sophmore-level computer organization course based on Patterson and Hennessy's "Computer Organization" book.
Which are boring. (Similarly boring as the dozens of "how do I construct a DFA for this stupid regular language?" questions we get at the beginning of every semester.)
@WanderingLogic Oh, sure. I missed the salvagable part of the question, you pointed it out to me, I agreed. On my end, there is no further need for discussion about this question.
I'm not sure I agree with your sentiment, @WanderingLogic. When I look at some tags at arXiv, I see indeed lots of papers along the lines of "How to do X with technology Y." I don't think that that is computer science research, so I would be careful to adapt our scope to what I view as the academic field drifting apart (in my terms, into CS and Systems Engineering).
So yes, my view is that applied systems questions should be OT on Computer Science. (And don't be fooled, interesting questions remain.)
That doesn't mean it has to be the community consensus.
@Raphael Yes, I know interesting questions remain.
@Raphael I'm hoping that the community can come to the consensus that if it would be considered on topic for an ACM conference or journal, or an IEEE computer society conference or journal then it is obviously on topic for cs.se.
@WanderingLogic I don't know if this is per se "correct". My personal view extends, of course, to such conferences and journals.
"How to implement algorithm X with Hadoop" may be an interesting question, but it is not a CS question. The prevalence of such titles in some fields only underlines that there is lots of CS to be done (finding and understanding suitable models) but does not show that implementation is CS. Imho.
(I don't care too much about the terms. We may end up with Mathematics - Comput(er|ing) Science - Systems Engineering or Mathematics - [Structural Science of Computation] - Computer Science. I prefer the former, mainly because the nomer "science" is too much of a stretch for the engineering parts. And I don't have a good name for the structural science that remains after you cut away engineering.)
@Raphael I made a statement about something I hope for. I can attest that it is "correct" in the sense that I really hope for it. Otherwise I'm not understanding what you are trying to say.
@WanderingLogic I guess I meant that when "blindly" importing scope from elsewhere, you also import its scoping problems.
And that these problems are apparent in that CS struggles to contain mathematical/structural parts, scientific parts and engineering parts under one roof. We can follow suit, but I think we should not (necessarily). Imho, CS will split apart over the course of the next few decades. But, as I said, all of that is just my opinion/prognosis and need not inform site policy.
Currently, we allow (software) engineering questions. However, few fit this criterion; most are either structural/fundamental or programming (the craft).
Regarding hardware, I'm quite fuzzy about where principles of hardware architecture end and implementation resp. electrical engineering start. (The question I pinged you with out me off by starting with "Intel" and "AMD" in one sentence.)
> So: the right size might be somewhere around the size of a university department. Somehow, the cultural anthropologists don’t mind sharing a building with the physical anthropologists, and when they both find themselves at the Yale-Harvard football game, you can bet that they’ll sit together and find something anthropological to talk about.
> Similarly, at Stack Overflow, the Java Entity Bean programmers at insurance companies don’t mind all the iPhone developers asking Objective C questions about the horrible, horrible game they’re working on. Heck, they might become iPhone developers one day. And they both share the humiliation of not being able to fix their uncle’s virus-infested Windows XP machine when they’re home for Thanksgiving.
I can tell you from personal experience that students in computer architecture type "cs.illinois.edu" into their web browser, not "ece.illinois.edu" when they are looking for research in computer architecture.
At just about every faculty meeting we'd have the same d*** discussion: "how can we get students to realize that there is great computer architecture research (resp. OS research, networking research) going on in the ece department, and apply to grad school here."
The answer "you can't. The name for what we do is `computer science.'"
As I said, I don't know who'll end up on which side of the fence (border areas are gray, similarly to mathematics-TCS), but my observation is that methods and goals differ so much that the split is inevitable.
@WanderingLogic From what I know about the history in Germany, that was mainly rooted in which department contained the most proto-computer-scientists when the discipline was emerging (and still mostly undefined). In some places, it still informs the focus of the CS department.
As an organisation, they may be able to just cover all branches a little while longer, until it becomes clearer which branches can stand as their own discipline. Software Engineering definitely will (imho), from your experience is seems to be less clear with Computer Engineering.
I'm too much a theoretician to predict where exactly the non-theory CS and SW/C-E subfields will draw the line. All I'm saying is that they will draw it.
@WanderingLogic Trajectory of the field is not the same as policy of an organisation.
My opinion is mostly informed by the internal struggles we have in our own department which (as I hear from basically everybody in the field I talk to) are echoed in many places. Engineers view things fundamentally differently from Theoreticians, and Applied Scientists fall somewhere in between. It's quite instructional to compare the inner workings of a typical CS department (which includes TCS, "core" CS, applied CS and SW/Comp eng) with, say, a mathematics department.
For instance, there is little dispute about what a mathematics student should be exposed to in their first year. In CS, that's a huge discussion, part of which stems from that we all don't really know what CS is, but parts also stem from very different and often orthogonal ideas for and requirements on the graduate.
I have found more and more that companies want SE but those programs don't really exist. SE is a lot more than learning how computers work and how to write basic software. Its a combination with an actual engineering process and to be honest software development IS an engineering process today.
I think the comparison with physics and mechanical/electrical engineering is quite accurate. (At least, as far as I can tell "from a distance"). Imagine these fields under one roof...
@baconoverlord Sorry, but as a researcher I couldn't care less what companies want. ;) That said, your observation is almost certainly a symptom of the same divide I'm observing/predicting: there is a need for a certain type of engineer but there is no branch in education/training that provides them.
@Raphael I can. Quite easily. University of Illinois had one of the last two "Theoretical and Applied Mechanics" depts in the U.S. until about 5 or 10 years ago. The dean merged them into the MechE dept. A few chose the Physics dept instead. (Which at Illinois is in the college of engineering for historical reasons.)
@baconoverlord Compare electrical and mechanical engineering: both contain some physics and some maths, and then they diverge. Both physics and mathematics continue to exist with their very own problem sets.
@WanderingLogic And how does that work out for them? (Still, they did not merge with physics, which would be the thing I wanted to suppose.)
(Over here, our architecture-construction engineering-regional and environmental planning department split into three only recently. But that's too far off our soup, I think.)
@Raphael Actually it is much more interesting than that. Many of the solid state physicists who work on electronic devices consider themselves "solid state physicists", belong to the APS, and wouldn't dream of leaving the EE dept they are in.
@WanderingLogic That may be for historical reasons, or funding reasons. They still say they are physicists. We have similar effects in TCS: many mathematicians get CS funding for doing maths, and quite a number of CSists venture into essentially pure mathematics.
Overlap areas exists and will continue to do so, but that does not prevent the split of the cores.
Is it as simple as a split between Engineering and Science? Engineering being the necessary tools to build and design things to a specific quality metric as part of a team or the Science to design things based in theory. This makes me worried sometimes that we will get graduate students with even less functional knowledge than we currently get.
Anyway, time will tell. For what it's worth, a leading Software Engineer (Dieter Rombach) who lectures and my university agrees with me. (Yay for arbitrary authorative reference hammer!)
What functional knowledge does a mathematics, astronomy or history PhD student have? None, because that's not their job and not their qualification. Imho, the idea that universities should train people for jobs is one of the banes of the current education systems. Companies should train their workers in apprenticeships, universities should train researchers in their programs.
A small overlap is well and fine, but moving job training to universities completely (for some job profiles) is debilitating.
@baconoverlord In the above discussion, I was mainly concerned with the research areas. Students are another issue, and the discussion starts with what "better equipped" means and for what you want to be equipping them.
In CS and as a researcher, yes, dropping all the engineering stuff would give us more room to focus on things important for CS researchers. I know engineers feel the same the other way around.
If that's better, I don't know. We might see a violent split and then a subsequent drift together (for a little bit) once people realise that they need each other after all.
(Regarding job training at universities, my boss put is quite aptly once: how can they possibly expect professors to provide job training? These are people that have never had a regular job in their lives! (Usually.))
@baconoverlord Yea, companies do that over here, too. Utter idiocy, imho. The reasoning seems to be: our best people are academics. So we need more academics. So let's shape education so that 1) the topics fit our needs and 2) we get more graduates. They completely forget that they change the game and hence the human material they get out. (This seems to be quite apparent with economics graduates over here.)
Next step: huh, our best people are PhDs. So let's get more of those and make sure that they work on the stuff we want from them, anyway!
And so on.... repeat until the whole higher ed system is geared towards providing "optimal" job training. Then wonder where science has gone.
Seriously, it's gotten so bad that students get really worked up because we require them to have independent, original thoughts in exams. They assume that they should pass when they have rote-learned "enough".
@baconoverlord It's similar here, mainly because of funding pressure. So far, at least at my department, professors remain conservative enough not to change anything in too big ways. Standards drop in individual courses, though.
Our engineering departments actually started a political war with the maths department because "too many" students failed their maths course. Now they pass these with barely more material than what is in the A-levels, and without any rigor.
I guess science departments can do what they want: their failure might hinder innovation, but no biggy. Incompetent engineers, however, are a danger to society.
I shiver in fear every time they proudly say that this or that essential feature in cars or planes is now controlled by software. Do you guys know about the Eurofighter? It can not stay airborne without computers. Those pilots are crazy.
@Raphael Indeed. Scariest talk I ever saw was by one of the guys who had designed the testing procedures for the Boeing 777 (Boeing's first fly-by-wire plane).
So, gotta go and finish of a useless day of brainwork with some honest legwork. Always appreciate the discussion and banter, guys! See you later/next time.
@WanderingLogic What's that in percent of total transit time?
I am self learning formal methods. I heard that formal methods is used (and usually only used) to create mission critical software (such as nuclear reactor controller, aircraft flight controller, space probe controller). That's why I am interested to learn it :p
However, after learning formal me...
