tag:blogger.com,1999:blog-6454006.post4874006895361435003..comments2024-01-16T14:32:49.175+00:00Comments on Arcane Sentiment: Quantitative results are rewardingArcane Sentimenthttp://www.blogger.com/profile/04144052171693893368noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-6454006.post-5385270888050424092012-05-10T03:32:13.466+00:002012-05-10T03:32:13.466+00:00It's less impressive if you know how horrible ...It's less impressive if you know how horrible the original implementation was, and that neither the 10x improvement nor the 3x involved any great cleverness. I just figured out what was making it slow and applied mundane solutions. I can't help thinking that a better programmer could have figured it out faster (and implemented parts of it in a less bug-prone way), or that it could be another order of magnitude faster if only I were a little cleverer.<br /><br />But the real point here isn't that I felt good because I did something impressive. It's that I felt good even without doing anything impressive, simply because I had a number telling me I had accomplished something.Arcane Sentimenthttps://www.blogger.com/profile/04144052171693893368noreply@blogger.comtag:blogger.com,1999:blog-6454006.post-9029767114048784742012-05-09T22:34:47.832+00:002012-05-09T22:34:47.832+00:00Okay, I'm duly impressed.Okay, I'm duly impressed.John Cowanhttps://www.blogger.com/profile/11452247999156925669noreply@blogger.comtag:blogger.com,1999:blog-6454006.post-84050783704741983142012-05-02T01:59:57.609+00:002012-05-02T01:59:57.609+00:00I got an order-of-magnitude speedup earlier. Addit...I got an order-of-magnitude speedup earlier. Additional improvements were harder to find (and much needed), so 3x still felt like an accomplishment.Arcane Sentimenthttps://www.blogger.com/profile/04144052171693893368noreply@blogger.comtag:blogger.com,1999:blog-6454006.post-66149274351637755092012-04-29T02:24:10.124+00:002012-04-29T02:24:10.124+00:00I decided a long time ago I was only going to be i...I decided a long time ago I was only going to be impressed by order-of-magnitude speedups.John Cowanhttps://www.blogger.com/profile/11452247999156925669noreply@blogger.comtag:blogger.com,1999:blog-6454006.post-89199410017725878032011-04-27T20:44:26.041+00:002011-04-27T20:44:26.041+00:00I use such a tool for testing Python code. I have...I use such a tool for testing Python code. I haven't noticed it having any effect on my motivation to fix bugs. But this is probably because a "bug" (as the concept manifests in my own thinking) is always something that is reported by a user, that our tests did not catch -- because we did not expect it.<br /><br />I suppose it increases my motivation to write tests for my code somewhat, but that's almost circular -- I could spend all my time writing testing all kinds of cases, likely or not, just to get more passing dots! But I don't, partly because it doesn't feel like it's going to help catch any real bugs (you know, conceptual shortcomings, race conditions, etc.), partly because it feels like the tests are largely there to compensate for the lack of static analysis in the language itself, and partly just because writing tests is <i>so boring</i>.That Bassett Disasterhttps://www.blogger.com/profile/17875046009465426632noreply@blogger.com