Episode 75 - A Drawer Full of Twixes (Miguel de Icaza ...)

[Fab]

On this special episode, we talk to Miguel de Icaza of Novell (co-founder of Gnome) about his work on the Mono and Moonlight projects, what he thinks about the Novell/Microsoft deal and the surrounding controversy, programming technologies in general and much more.

http://linuxoutlaws.com/podcast/75
http://linuxoutlaws.com/podcast/ogg/75

http://outlawarchives.com

[mynameis]

I am waiting patiently. (beer isn't cold yet, anyway)

[Fab]

It's out!

[NYbill]

i think he is waiting until the beer is cold to listen. after all, a nice beer and a LO episode go quite well together.

downloading....

[Fab]

i think he is waiting until the beer is cold to listen. after all, a nice beer and a LO episode go quite well together.

Exactly! That's the spirit! Makes the ranting come easier as well....

[emma]

great show, just finished listening to the interview and really liked it.

[gorem]

Listening now.......

[mynameis]

That was far too pro M$... I shall go and wash.

Great song BTW.

[gorem]

Excellent show and I'm glad I've heard from Miguel.

Far from being pro-MS in my opinion he comes off in this interview as being realistic. This technology is being used, get over it........

The best approach to challenging MS and its pervasive and corrosive use of software patents is to 'push back' by using its technology in ways that renders them if not irrelevant, definitely pointless.

Any exposure Windows developers get to Linux and FOSS in a wider context is good.

Miguel's point about Mono being used for creating solutions/apps that no-one wants to write is a good place for mono to be in as an adoption strategy for widespread acceptance of Linux too......

MS though is a corporation and is capable of anything.....

Oh and another vote for a cool choice of music.

[soyrochus]

Good show guys. Informative. Balanced-but-not-Boring. And rather refreshing to hear something more varied then just the average "M$ == evil" rant. I lik(ed) the idea of Mono as it promises to make the creation of true multi-platform applications a bit easier. However, like with the interview with LUG Radio, De Icaza seems to be less forthright, if not a bit disingenuous, then he appears to be at first sight. The issues surrounding the Novel / MS deal cannot be so easily whisked away. Look at this excellent resource page at Groklaw: http://www.groklaw.net/staticpages/inde ... 8045851480

There are serious issues surrounding the emulation of the non-standard (non ECMA) elements of the DOT.NET framework. Now, rationally speaking, I would agree with De Icaza that the real risk of Microsoft heaving this axe is rather remote. There are so many more ways in which Microsoft can royally screw the F/OSS community. And Microsoft *has* been rather good in releasing a series of products under *real* OSS licenses. But then again, Microsoft is not one single monolithic entity, especially since the departure of His Billness. So you never know what shit the bowels of the company might release on us, even if the head seems to be a prim and proper. I guess in the end the real risk comes from the fact that no-one is looking at this "rationally", certainly not the market or the average IT "manager" (haha). So any threat of litigation, however remote, could hinder adoption of OSS, which would not hurt Microsoft. So in that sense, a strong presence of Mono on Linux in general *is* a potential risk.

I also couldn't shake the impression that he was playing a bit too much of the sales manager in the way he was dissing Java and praising Mono. Certainly Mono is impressive. But recently I had the opportunity to look at Mono 2.0 and 2.2 extensively and there are a significant number of issues which make it far less attractive even from a technical point of view. In short: "base" Mono, the combination of CIL VM/compiler, C# language implementation and class libraries, is very well done. However, the tool chain surrounding it is buggy and often rather crappy. Something as simple as generating a source file from an WSDL, which in 2009 really is no longer a feat of magic, will on Mono result in many failures. The Mono debugger was delivered only recently and is itself chock-full of bugs. And the IDE, MonoDevelop, which De Icaza praised is in version 1 a glorified editor. The current alpha of version 2 shows promise but for the moment, though it has more features, it still pales in comparison with many other IDE's (and it's unstable as hell, but ok, it's alpha software). Worst of all is that Mono is a real bitch to compile and not well supported on platforms other than SUSE (Novell). Ever tried to get Mono 2.2 build on Ubuntu 8.10?

That all is a severe problem for Windows DOT.NET developers which are supposed to be tempted to cross the divide. Mono just doesn't live up to what they expect. Java, or rather, the Java Virtual Machine, does. It offers many advantages; it runs on nearly every imaginable platform, is stable, fast, comes with excellent tools, an enormous installed base, strong industry support, the best IDE's on the market and a huge Open Source community to boot. This community easily dwarfs that surrounding DOT.NET and Mono combined. And, like DOT.NET, you are not bound to a single language on the JVM either. There are stable implementations of other languages: Scala, Groovy, Python ("Jython") and Ruby ("JRuby"). And we are not just talking Java here: the *nix, C based, ecosystem (Apache, Python, KDE, Gnome, what not) is nowadays at least as advanced as either Java or DOT.NET.

So, recently, after struggling with several serious issues, I've opted to leave Mono for what it is and we will be using Python for a large project. Nowadays it comes with terrific IDE's (based on Eclipse or Netbeans, but "native" python IDE's like Eric are excellent as well). And with ctypes you don't have to be a C wizard to use the plethora of available C libraries. You don't notice that gPodder is written in Python, do you?

Anyway, I will definitely keep an eye out for Mono though. For better or worse