RE: [Python-Dev] Looking for master thesis ideas involving Python
From: Dennis Allison [mailto:allison@sumeru.stanford.EDU]
How about re-engineering the interpreter to make it more MP friendly? (This is probably a bigger task than a Masters thesis.) The current interpreter serializes on the global interpreter lock (GIL) and blocks everything.
To me this would probably be the most interesting thing to tackle - especially since it has been tried before with partial success but overall failure. At the very least that gives a body of work which you can refer to both as a starting point for your work, and to show how your approach differs from and improves on existing work. It would also be of tremendous value to Python IMO if it could be done without negatively impacting performance on single-processor machines. Whether it is too large for a Masters thesis I don't know. Does a Masters thesis require *success* in the stated goal? I've been thinking about doing my own Masters in the not-too-distant future if I can find the time ... Tim Delaney
Measuring the size of a project is difficult. This one would require (I think) some significant out-of-the-box thinking. There are a number of resource which could be brought to bear in addition to Herlihy's work on synchronization, for example, Kourosh Gharachorloo's work on programming for the Stanford Dash MP where he toyed with the issues involved with building synchronization independent (that is, lock independent) programs. On Thu, 30 Oct 2003, Delaney, Timothy C (Timothy) wrote:
From: Dennis Allison [mailto:allison@sumeru.stanford.EDU]
How about re-engineering the interpreter to make it more MP friendly? (This is probably a bigger task than a Masters thesis.) The current interpreter serializes on the global interpreter lock (GIL) and blocks everything.
To me this would probably be the most interesting thing to tackle - especially since it has been tried before with partial success but overall failure. At the very least that gives a body of work which you can refer to both as a starting point for your work, and to show how your approach differs from and improves on existing work.
It would also be of tremendous value to Python IMO if it could be done without negatively impacting performance on single-processor machines.
Whether it is too large for a Masters thesis I don't know. Does a Masters thesis require *success* in the stated goal? I've been thinking about doing my own Masters in the not-too-distant future if I can find the time ...
Tim Delaney
_______________________________________________ Python-Dev mailing list Python-Dev@python.org http://mail.python.org/mailman/listinfo/python-dev Unsubscribe: http://mail.python.org/mailman/options/python-dev/allison%40sumeru.stanford....
Delaney, Timothy C (Timothy) wrote:
From: Dennis Allison [mailto:allison@sumeru.stanford.EDU]
How about re-engineering the interpreter to make it more MP friendly? (This is probably a bigger task than a Masters thesis.) The current interpreter serializes on the global interpreter lock (GIL) and blocks everything.
To me this would probably be the most interesting thing to tackle - especially since it has been tried before with partial success but overall failure. At the very least that gives a body of work which you can refer to both as a starting point for your work, and to show how your approach differs from and improves on existing work.
It would also be of tremendous value to Python IMO if it could be done without negatively impacting performance on single-processor machines.
Whether it is too large for a Masters thesis I don't know. Does a Masters thesis require *success* in the stated goal? I've been thinking about doing my own Masters in the not-too-distant future if I can find the time ...
Success as in what you set out to do was actually beneficial? No, just as long as something is learned. Successful as actually finishing the darn thing? Yes. Basically a masters thesis needs to require some research, such as looking at other implementations, and some original thought if possible. The problem with a masters thesis, though, is that I have a fixed timeframe (want this done in about a year's time for doctoral school applications) and I don't get to spend a large portion of my time on it (I still have to take normal classes during this time, although I can fenagle my schedule to minimize my work load). I will still consider this, though. -Brett
Brett C. wrote:
Whether it is too large for a Masters thesis I don't know. Does a Masters thesis require *success* in the stated goal? I've been thinking about doing my own Masters in the not-too-distant future if I can find the time ...
Success as in what you set out to do was actually beneficial? No, just as long as something is learned. Successful as actually finishing the darn thing? Yes.
He actually meant "success in the stated goal". I.e. if you go out to implement free threading, would it be considered as a failure of the Master's project if you come back and say: "I did not actually do that"? My answer is "it depends": If you did not do that, and, for example, explain why it *can't* be done, than this is a good thesis, provided you give qualified scientific rationale for why it can't be done. If you say you did not do it, but it could be done in this and that way if you had 50 person years available, then this could be a good thesis as well, provided the strategy you outline, and the rationale for computing the 50 person years is convincing. If you just say, "Oops, I did not finish it because it is too much work", then this would be a bad thesis. Regards, Martin
Martin v. Löwis wrote:
Brett C. wrote:
Whether it is too large for a Masters thesis I don't know. Does a Masters thesis require *success* in the stated goal? I've been thinking about doing my own Masters in the not-too-distant future if I can find the time ...
Success as in what you set out to do was actually beneficial? No, just as long as something is learned. Successful as actually finishing the darn thing? Yes.
He actually meant "success in the stated goal". I.e. if you go out to implement free threading, would it be considered as a failure of the Master's project if you come back and say: "I did not actually do that"?
Ah, OK. My mistake.
My answer is "it depends": If you did not do that, and, for example, explain why it *can't* be done, than this is a good thesis, provided you give qualified scientific rationale for why it can't be done. If you say you did not do it, but it could be done in this and that way if you had 50 person years available, then this could be a good thesis as well, provided the strategy you outline, and the rationale for computing the 50 person years is convincing. If you just say, "Oops, I did not finish it because it is too much work", then this would be a bad thesis.
I would have to agree with that assessment. Just have to convince my thesis adviser. =) -Brett
How about re-engineering the interpreter to make it more MP friendly? (This is probably a bigger task than a Masters thesis.) The current interpreter serializes on the global interpreter lock (GIL) and blocks everything. ... I will still consider this, though.
-Brett
If you take this on there is no doubt you'll receive many-a-beer from people on this list! :)
participants (5)
-
"Martin v. Löwis" -
Brett C. -
Delaney, Timothy C (Timothy) -
Dennis Allison -
Gregory P. Smith