Mono is fast!
|
| Author | Content |
|---|---|
| bigg Jun 12, 2009 6:53 AM EDT |
This was only a minor point at the beginning, but it really hurt the credibility of the author in my eyes:Quoting:Speedwise, Mono is much faster than Python - up to several hundred times faster according to some benchmarks. I'm not even going to bother explaining why that is at best misleading. This individual prefers Mono. That's fine - but say so, that it's a personal preference, most likely based on what was learned in school or on the job. Most programming is not done in Mono and I doubt that is because of ignorance of what Mono can do. |
| KernelShepard Jun 12, 2009 7:13 AM EDT |
My guess is that he was referring to the Debian Language Shootout benchmarks: http://shootout.alioth.debian.org/u32q/benchmark.php?test=al... As you can see there, Mono is 276 times faster than python for the spectral-norm benchmark test and 95 times faster than python for the fannkuch test. Quite a bit faster in a number of other tests too. Of course Mono also seems to be half as fast as python for the Regex test. |
| bigg Jun 12, 2009 7:30 AM EDT |
But as you and pretty much everyone knows, you don't take a single test and use that number. I've seen tests for which Java did better than C, but you don't even need to look at the data to know that's nonsense. The link you've given refers to numerical analysis. There is cython if you want speed (I'm guessing that's a feature Mono doesn't have, and would in almost all cases be faster than Mono), pygsl offering libraries for numerical computing, rpy2 to trivially connect to R, and f2py to put a trivially easy interface on Fortran code, just to name a few options. It makes no sense at all to run numerical code in native Python if you are after speed. But yes, you can always find something that one language does better. The story can be summarized as follows: there is no patent threat and developers have a preference for Mono. That's fine. I just don't appreciate FUD about other languages as justification. |
| KernelShepard Jun 12, 2009 8:48 AM EDT |
I'm not a big believer in benchmarks as they can obviously only show a small subset of things a programmer might want to do. Also, I have my doubts about the Debian Language Shootout in particular since I've personally discovered that a number of the programs in various languages to be grossly inefficiently implemented in comparison to their counterparts. That said, I was just pointing out where his numbers seemingly came from (he might be referring to other benchmarks, I have no idea - but having seen the Debian benchmarks and seeing that he's a Debian developer, it seemed he might be referring to those). Yea yea, assumptions make an ass out of you and me :-) |
| Sander_Marechal Jun 12, 2009 9:38 AM EDT |
Quoting:Also, I have my doubts about the Debian Language Shootout in particular since I've personally discovered that a number of the programs in various languages to be grossly inefficiently implemented in comparison to their counterparts. I've seen that too. People using C or Java style loops in Python instead of list comprehensions (which are much faster of course). |
| kingttx Jun 12, 2009 2:51 PM EDT |
For me and many others, I think the speed factor is not the immediate worry. If Mono is fast, cool beans! If it fosters some really good apps and helps developers port .NET code over to Linux, nice! What isn't being answered is whether the terms of the technology Mono is based upon are 100% Free, no restrictions limiting the use to non-commercial installs. Also, no one is saying outright that Mono is relatively and legally "safe" to use and redistribute, at least no one that bears the authority to do so. Saying we are covered by OIN isn't the point. This isn't a nuclear arms race, this is trying to keep the FOSS environment free/Free and clear so we don't naively hand someone an easy route to begin doling out expensive lawsuits, even if their grounds to sue aren't entirely genuine. Leaning too heavily on OIN protection may spend our leverage out too soon. IS MONO FREE? Once that gets clearly and authoritatively answered, then we can move on to debating its technical merits. In the meantime, anything else ("best of breed", "it's FAST", etc.) is absolute noise. |
| bigg Jun 12, 2009 4:13 PM EDT |
@kingttx: That's the million dollar question. Advocates of mono ask for evidence that it's not free, while critics of mono ask for evidence that it is free. I don't see that the essay does anything for the debate. (And I remain indifferent because I still don't see a reason to use Mono.) |
| dumper4311 Jun 12, 2009 5:33 PM EDT |
@bigg: An eloquent response to what could easily (and may still I guess) escalate into yet another round of flogging a deceased equine. While I don't personally have any use for Mono either, I wish it's supporters good fortune. May they happily scratch whatever itch they are reaching for - that is after all the point of this much vaunted freedom. |
| caitlyn Jun 12, 2009 5:41 PM EDT |
My only concern, if it isn't free, is that it doesn't become so enmeshed that it is used to undo whole valuable projects. Microsoft wants Linux and FOSS in general destroyed. Therefore it is incredlibly hard to trust anything that comes from them. This isn't about "not invented here". It's about whether or not it has been invented to serve as a time bomb waiting to go off. |
| dumper4311 Jun 12, 2009 5:55 PM EDT |
Let's pretend that is the case, and that Mono is a spectacularly successful time bomb. So what? Legal issues have been coded around before, and there's no capability in Mono or .Net (that I'm aware of) that couldn't be argued around via prior art or various other methods if necessary. If it blows up, and if all these projects are rendered inert, other projects will spring up to replace them. Again, the benefits of F/OSS development. Will it be inconvenient for some? Probably, but life goes on. Making this issue such a focal point is counter productive. |
| kingttx Jun 12, 2009 5:57 PM EDT |
@bigg: That appears to be one of the advocates' many "tools" for Mono, "Why don't you do some research and see if there are any patents that Mono may infringe upon?" Well, that's not the responsibility of anyone OUTSIDE of the project. In fact, I'd venture a guess that Mono advocates have no authority to state unequivocally that Mono is indeed safely and 100% Free. Pressing many of them to point to such a formal statement usually brings responses similar to "Well, other software just might infringe on patents, too!" It's like chomping down on a burger only to have the tomato fall out from the bottom. 1. We aren't talking about other projects, we're talking about Mono. Don't change the subject! 2. There are few application projects that purposely build upon such dubious terms like .NET's. From what I gather, the RAND agreement behind this technology is not 100% Free, but I may very well be mistaken. 3. Just because something is a standard doesn't mean a patent holder won't sue. (SDRAM memory, anyone? Rambus wants its share!) I agree, bigg, that the essay isn't answering the main question. Oy! |
| tuxchick Jun 12, 2009 5:58 PM EDT |
What set off this latest round of Yay Mono/Die Mono follies is Ubuntu including some Mono apps by default in the installation image. It is seen as the beginning of a long-term strategy to push Mono firmly into Linux whether users want it or not. It seems that the majority are Ok with Mono being included in the standard repos, but think it's pushy to force it into the default. And now we see that Debian Squeeze will also include Tomboy by default, http://robertmh.wordpress.com/2009/06/12/mono-in-the-default... The infamous Josselin Mouette http://np237.livejournal.com/ is part of this; he claims that he had no part in the decision, but he sure does a number on the people who don't like it. With these kinds of attitudes, language, and tactics, it's easy to believe that there is a sinister Mono agenda at work. They keep bleating about "best of breed" as though it were carved in stone, when actually it's debatable. Sure does remind me of classic Sauron tactics. |
| dumper4311 Jun 12, 2009 5:58 PM EDT |
Where's tracyanne at? If I recall, she's our resident Mono expert, and I'd be interested to hear what she has to say about the contents of this article. The original article mind you, not so much the "is mono free" question, we've all gone the rounds on that before. . . . |
| tuxchick Jun 12, 2009 6:07 PM EDT |
Regarding patents, I thought this comment to Jo Shield's article was pretty good. Nice and short, anyway: "Grr.. said: Come on! We can't have stupid mp3 support by default because of possible patent violation, but it's ok to have by default whole armada of MS patents and build our userland on top of them? Just who is crazy one here? June 12, 2009 5:07 PM" |
| azerthoth Jun 12, 2009 6:12 PM EDT |
"I wont use mono because I cant be certain it's free" lame better stop using Samba too, or better yet, linux entirely, it has been said yet unproven that linux has several hundred violations. They have been laid out with all the proof that mono has. I'm sorry, maybe's and mights, if they are your thing, if you look hard enough you will find contrary maybe's to just about everything. It just doesnt hold any water. I can say that I dont use mono, not from a philosophical view, but because I haven't found an app that requires it, that I dont think is a complete waste of resources. |
| Sander_Marechal Jun 12, 2009 6:51 PM EDT |
Quoting:Microsoft wants Linux and FOSS in general destroyed. Linux, yes. But not FOSS in general anymore I think. Their strategy seems to be to get all FOSS on Windows. Mono firs into that. It makes it trivial to port Linux/Mono applications to Windows/.Net. It tries to ensure that there can be no killer app exclusive to Linux. I'm not sure that a Mono patent trap fits into that strategy. On the one hand they could first pick up all the great Linux/Mono apps, put them on Windows and then spring a trap, hoping that the apps stay on Wiindows instead of Linux. On the other hand, since most of the developers of such apps are on Linux, having a Mono trap would just ensure that the new versions of those apps aren't coming to Windows anymore. It would be counter to their strategy. |
| igouy Jun 15, 2009 8:13 PM EDT |
Quoting:KernelShepard > I've personally discovered that a number of the programs in various languages to be grossly inefficiently implemented in comparison to their counterparts So contribute better programs! http://shootout.alioth.debian.org/u32q/faq.php#play |
| igouy Jun 15, 2009 8:15 PM EDT |
Quoting:Sander_Marechal > People using C or Java style loops in Python instead of list comprehensions (which are much faster of course). Easy to say, show it! Contribute better programs. http://shootout.alioth.debian.org/u32q/faq.php#play |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!