tag:blogger.com,1999:blog-6555947.post109529027307627733..comments2024-03-14T01:32:43.610-06:00Comments on The Geomblog: Experimental Work in SODASuresh Venkatasubramanianhttp://www.blogger.com/profile/15898357513326041822noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-6555947.post-1095555070642028772004-09-18T18:51:00.000-06:002004-09-18T18:51:00.000-06:00Suresh,
I think the link to Johnson's paper is br...Suresh,<br /><br />I think the link to Johnson's paper is broken.<br />It should be<br />http://www.research.att.com/~dsj/papers/experguide.pdf.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1095553399099033302004-09-18T18:23:00.000-06:002004-09-18T18:23:00.000-06:001. It is not clear that there is a paucity of expe...1. It is not clear that there is a paucity of experimental papers overall: ALENEX has been attracting attention and both ESA and SoCG have applied tracks. <br /><br />2. There is no reason to expect a large number of experimental papers at SODA, similarly as there is no reason to expect a large number of (say) CG papers at FOCS. Some good ones will get into SODA, and that is fine. There are other avenues, and ALENEX is a good one. <br /><br />3. It is not clear to me that "doing experimental work" is detrimental from a "career" point of view. It all depends on how one frames one's work, and ultimately, you do what interests you. If one were to say that only a small fraction of the theory community is interested in experimental work, that is probably true, and I see no problem with it. However, for those who do find experimental work interesting, they will find ways of placing their work in context so that it doesn't harm their career.Suresh Venkatasubramanianhttps://www.blogger.com/profile/15898357513326041822noreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1095520732689764132004-09-18T09:18:00.000-06:002004-09-18T09:18:00.000-06:00I think a horse race paper is quite useful. Perhap...I think a horse race paper is quite useful. Perhaps<br />the most important side-effect is that code for<br />various algorithms is actually written and tested<br />which happens too infrequently in the algorithms<br />community. Asking for much more than this is hardly<br />reaslistic given the amount of work involved. <br /><br />At the same time, I am not convinced that the current<br />SODA model where both theory papers and experimental<br />papers are evaluated side by side is going to work<br />unless there is some explicit quota for experimental<br />papers which is also problematic. Basically I don't<br />think there is any incentive, career wise, for the theory/algorithms community to do experimental work<br />and that more than anything else explains the paucity<br />of papers in the area. I don't think this is a bad<br />thing in itself - perhaps there are not enough applications today that truly need all the sophisticated algorithms developed in theory so we'll have to just wait for that to happen. <br />In the mean time if there are too many theory people the market should take care of that eventually.<br /><br />ChandraAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6555947.post-1095517988727731502004-09-18T08:33:00.000-06:002004-09-18T08:33:00.000-06:00I do not religiously read the SODA proceedings, so...I do not religiously read the SODA proceedings, so I probably have missed some great works. However, the only experimental SODA paper I have admired was: <br /><br />J. Kleinberg. Authoritative sources in a hyperlinked environment. Proc. 9th ACM-SIAM Symposium on Discrete Algorithms, 1998.<br /><br /><br />That is a great piece of experimental work.Anonymousnoreply@blogger.com