Interview (in English): Miguel de Icaza - I (Mono)

13. 6. 2005 | Filip Vrlík | Rozhovory | 12328×

1. Filip Vrlík: On which projects do you focus mostly at this time? Can you introduce these projects to us, please?

   Miguel de Icaza: I work mostly on the Mono project. Which is a developer platform, based on the .NET developer platform. What is interesting about this technology is that it is independ on the language. Many people have spend enormous time by making sure that this virtual machine is good enough...

2. Something like Java?

   It's similar to Java, but they have put an extra emphasis on multiple language support. For example, one thing that Java can't do efficiently is to support C and C++.

   Motivation for Mono is productivity. We have built Evolution and OO.org in the past with C++. And a lot of the problems that programmers face today are, ehm, C++ is very low-level language. You have to do your own memory management, so your application leak if you are not carefull. Also there is a lot of problems with memory coruption; like you free a pointer twice, you use the buffer that someone else is using, or you corrupt memory that someone else is using, and so on.

   We were looking, or Ximian was, for a platform that would be useful for creating large applications like Evolution in short amount of time or with less people. Either with less people or create more. Instead of creating Evolution that took two years to develop, they can do it in one year. We looked at Java and C# and it seemed that C# has the same as Java plus something more. Whenever you are doing something, whatever it is, time is important. So we have done it this way.

   As you can see, Mono was a technology to make things on the desktop happen. I'm most interested on Linux on the desktop, so, ehm, the reality is that I'm still developing for the desktop, but Mono means for any area. I'm still working with desktop group. We had just huge conference recently in Germany, I'm coming from there. They are not directly hacking desktop, but I'm providing tools for them. We have built GTK#, which is toolkit for building applications for Linux. We are making designer tool. We provide good developer tools for people and documentation. Many people complain that Open-Source documentation is terrible. We are working on documentation for Mono, we document everything in Mono, so people have good docs. We are trying to fill in the holes for thing needed to make linux successful on the desktop.

   (Miguel's mobile is ringing)

   That's my alarm. I was gonna wake up now, imagine that.

   Tell me little bit about your magazine.

   (I'm telling him about Abclinuxu.cz)

   I have another question. In each country is the situation with Linux different. How's Linux doing in Czech republic?

3. It's not so good, but the situation is going to be better in my opinion. The community of Linux it's not so huge as in other countries. But it isn't bad and it's going to be good...

   I will tell you a story.

   Some of the best programmers I have had chance to work with were from Czech Republic. I worked with many different ones. Coming from Mexico... or this way, if you are looking from the outside not from the inside, Czechs are on the top, percently twice more than the others contribute to Open-Source. Maybe it is because of education system.

4. I don't think so . But thanks.

   You have been talking about Mono. What are the news about Mono project? Any success recently?

   I think the most important thing is that... We released Mono last year, about ten months ago. And it's finally maturing to the point of being a solid code. We thought 1.0 will be never usable and 1.0 was usable, but it wasn't as polished as we have it today. We are polishing so much, every detail in Mono. So, you have better experience. Lots of polish. That means better performance, we recently have made a lot of work for performance. Some of the work we were doing will sadly not show up in 1.2, it's gonna show in the later version. Because of really big architectural changes, well not changes but some optimizations are very very hard. So, we are not gonna have it in September, as we supposed.

   Better performance, stability, right. Again it's a thing, I think, people don't really realize. We have learned from long running servers or long running applications on the desktop like Beagle desktop search. He was more stable than people think. Three big things, features, that people will notice. It's nice that you are stable and faster, but it doesn't really matter. But we are coming out with new GTK# which will suppport even more things in GTK. We have new tool to automaticaly wrap any GObject library, that you have written in C. And you can automaticaly use it from C#. We have developed some bindings to Evolution, you can completly script it with C#. I said C#, but it could be any language supported by .NET Python, Visual Basic, JavaScript or any other third party. So now you can control Evolution, you can script Evolution, use all the Evolution mail features.

5. I see you prefer C#. Right?

   I prefer C#, because I just use it. But other people in the team prefer Python or others. So, they use that. C# is lighter, lighter of them, because C# doesn't really have a C# run-time. C# directly communicates with libraries. Well, in the case of Python, it has a boot code. Java has run-time and then the class libraries. C# is the lightest of them. But it doesn't mean you should choose it for this reason. It would be stupid, choose it only for this reason. My preference, being an old man, is to choose the one I know.

6. People have strange feelings about Mono, they see Microsoft and it's technologies, problems with copyrights and so on...

   No, no, no! There are no copyright problems.

   So they see Microsoft and his technologies and they have strange feeling as I said. On the other hand it's good to have a backdoor or something in case of successful spreading of .NET.

   Yes, well, I'm working the successful spreading of the .NET on the Unix platform. Hmm, for example in eight months we realesed Mono they have created at least 40 desktop applications. At least we have succeeded in term how do we get programmers to starts writing desktop apps. Because it was hard in the past.

7. If you rewrite some application, for example Gimp to Mono. What will it bring? It will be only slower, isn't it truth?

   Ok, two things. First - I do not think we should rewrite software, because everytime you say "let's rewrite the software" you are embarking in multi-year project. And it doesn't matter what language you are translating in. If they want rewrite a Gimp. Mmm. Gimp is a project, that has been around since 1996, maybe. It has been already rewritten one time, in year 1997 or 1996, because it was originaly written for Motif and then he was rewritten for GTK. These same people built GTK. And... so we are talking about an application that had at least eight years to mature. So this is very matured application. I don't believe we should invest our time in rewriting working code.

   Ehm, if we need to, if we need to extend this application with better programming language, my suggestion is to embed Mono, right, Mono is embedable language like Python. So you can embed Mono, link it with app, and then you can write extra code in Mono, but you don't need to rewrite the original code. It's millions of lines to debug. So my suggestion is not to rewrite but to fix box here and to extend it here. (rich mimics). For example you can do scripting from C# and write new features from C#, but you keep the existing code base, right. I think that's the road to success. I really discourage people who want to rewrite things. There was a suggestion to rewrite Evolution in C#. I told them: "don't do it, that's not the point, use the bindings." The code is working, is debugged, is in use, is in production. Why are we going to waste years to rewrite it and get it right? We would make the same mistakes so why exchanging some box for other box. We are human, right? So the point is, I am not suggesting people rewriting, I am suggesting - if they write new application, new ones, write it in Mono. Or extend existing application with Mono.

       

Komentáře

[MaT]

Tak ta druhá část co měla být o Gnome se nějak ztratila... Místo toho na "page=2" sedí ta první část o mono, akorát že v angličtině...

[Robert Krátký]

Ta druhá část se nikam neztratila. Jen vyjde o něco později. Rozhovor byl docela dlouhý, a proto byl rozdělen do 4 částí. Trochu to v článku upravím, aby to nemátlo.

[Sigi]

Tady je jiny pohled na mono na linuxu

Why Mono is Currently An Unacceptable Risk

T.

[Radek Podgorny]

Hmmm, to same (nebo podobne) by se dalo namitnout i proti Jave a nejaky ten hi-lev jazyk proste potreba je. Tak ktery? Mne se vic libi C#

[Petr]

Osobne mi prisel zajimavy take tento clanek prinasejici dalsi pohled (IMHO rozumny). Tusim, ze byl kdysi ve zpravickach. Jedná se o MonoForFun.

Petr

[monak]

No, možná je dobře, že java má po mnoha letech plnohodnotnou konkurenci nejenom na platformě Windows.

Když Novell "koupil" mono, tak jsem si myslel, že má jediný cíl a to mono zrušit, ale hodně jsem se mýlil. Což mě těší.

Jinak mohu říct, že udělat formulář ve Visual Studiu, zkompilovat ho a tento mezikód spustit na linuxu je velice zajímavá věc. Takže pro Vaše info, aplikace napsaná čístě v C# půjde spustit v linuxu bez dalších kompilací.

Dále, když vezmu v potaz, že Microsoft chystá otevřený formát v XML pro další verzi ms-wordu, tak mono a .NET je v současné době hodně perspektivní. Bude to ideální platforma na tvorbu CRM aplikací.

Určitě bych mono v dnešní fázi nebral jako úlet, ale jako věc, na které bude možné stavět v blízké době firemní aplikace. Já určitě svoji .NET aplikaci po vydání nového office očistím od interop.

[aubi]

Pristi verze wordu s XML? Ze jsem tohle uz nekde slysel? Nerikali totez uz posledne? A taky o podpore HTML?

A s moznosti poustet veci zkompilovane pro widle na Linuxu: neni M$ prilis znamy tim, ze schvalne zavadi nekompatibilitu, aby jejich soft nechodil na jinem systemu? Uz se napravili?

Petr

[monak]

Co se týče toho formuláře, tak to jsem zkoušel. Takže to není teorie. Bohužel jsem to v příspěvku neuvedl. Ještě dodám, že mono vzniká po uvedení .NET 1.1 a ne současně. Takže pokud MS změní specifikaci, tak to odnesou i jejich aplikace. Což si asi nedovolí. Ale to je můj subjektivní dojem.

[aubi]

Presne totez, co s MS DOS vs. DR DOS -- Widle na DR DOSu proste nechodily. Vzdycky si dokazou najit tu kapicku nekompatibility. Je otazka, jak moc je to dulezity, protoze nekompatibilita Win2k se sambou taky moc dlouho netrvala. Z takovych triku mam strach.

Stejne tak moc neverim tomu s pravy. Nektere nezbytne casti jsou licencovane, takze cela aplikace spustit nepujde a prepisovat to nikdo nebude.

Je to porad dokola.

[orin]

Jinak mohu říct, že udělat formulář ve Visual Studiu, zkompilovat ho a tento mezikód spustit na linuxu je velice zajímavá věc. Takže pro Vaše info, aplikace napsaná čístě v C# půjde spustit v linuxu bez dalších kompilací.

Mozno sa mylim (skoro nic o tom neviem), ale niekde som cital, ze to zakazuje licencia k Visual Studiu.

[ND]

Podle mého názoru by to jen tak MS neprošlo, právníci budou znát institut "dobré víry", "veřejného příslibu", "nekalá hosp. soutěž" apod. V těchhle věcech se nevyznám, ale tak snadné by to MS asi neměl.

[aubi]

Ze tu otazku zamluvil? Necetl jsem angl. original, ale na primou otazku vubec neodpovedel.

Mne sice stve, ze bych mel pouzivat M$ technologii, ale nakonec, jestli je dobra, proc ne. Jenze co se stane, az se M$ rozhodne, ze chce za .NET platit???

Otazka zaznela, odpoved ne. Bude to v tech trech nasledujicich pokracovanich?

Petr

[llook]

Odpověď zazněla a velice přímá - nejsou problémy s autorskými právy. Asi předpokládal, že se všeobecně ví, že jde o ECMA standardy, tak si odpustil podrobnosti.

Jsou věci, které v Monu z autorskoprávních důvodů nebudou - třeba Windows Forms - ale samotné CLR a C# není problém.

[Michal Marek (twofish)]

V anglické verzi je

<p><i>That's my alarm. I was gonna wake up now, imagine that.</p>

A zbytek stránky je pak kurzívou.

[Robert Krátký]

Opraveno, dík.

[VícNežNic]

Skupina desktop je co? Předpokládám, že tou konferencí myslel GUADEC. Je Linux desktop a desktop Gnome? No nic

Tyhle vícestránkové články jsou novinka? Ještě jsem to tu neviděl.

[Robert Krátký]

Tyhle vícestránkové články jsou novinka? Ještě jsem to tu neviděl.

Není to novinka. Jen není tak často příležitost tuto funkci využít. Jestli si dobře vzpomínám, tak naposledy jsme ji použili v článku Co to je ABC Linux.