3 Q: Is Mono the same as Microsoft's .NET initiative?
7 .NET is a company-wide initiative at Microsoft that
8 encompasses many different areas. The .NET development framework,
9 Passport, Biztalk, new server products, and anything that is
10 remotely connected to .NET gets the ".NET-stamping" treatment.
11 Some components of Microsoft's .NET initiative have been announced
12 and some others are in the works.
14 Mono is a project to implement several technologies developed by
15 Microsoft that have now been submitted to the ECMA Standards Body.
17 Q: What technologies are included in Mono?
19 A: Mono contains a number of components useful for building new
23 * A Common Language Infrastructure (CLI) virtual
24 machine that contains a class loader, Just-in-time
25 compiler, and a garbage collecting runtime.
27 * A class library that can work with any language
28 which works on the CLR.
30 * A compiler for the C# language. In the future we
31 might work on other compilers that target the Common
34 Windows has compilers that target the
36 <a href="http://msdn.microsoft.com/net/thirdparty/default.asp#lang">a
37 number of languages:</a> Managed C++, Java Script,
38 Eiffel, Component Pascal, APL, Cobol, Oberon, Perl,
39 Python, Scheme, Smalltalk, Standard ML, Haskell,
43 The CLR and the Common Type System (CTS) enables applications and
44 libraries to be written in a number of languages. Classes and
45 methods created in one language can be used from a different
48 This means for example that if you define a class to do algebraic
49 manipulation in C#, that class can be reused from any other
50 language that supports the CLI. You could create a class in C#,
51 subclass it in C++ and instantiate it in an Eiffel program.
53 A single object system, threading system, class libraries, and
54 garbage collection system can be shared across all these languages.
56 Q: Where can I find the specification for these technologies?
58 A: You can find the work-in-progress documentation from the T3G ECMA
61 <a href="http://www.dotnetexperts.com">http://www.dotnetexperts.com</a>
63 Q: Will you implement the .NET Framework SDK class libraries?
65 A: Yes, we will be implementing the APIs of the .NET Framework SDK
68 Q: What does Mono stand for?
70 A: Mono is the word for `Monkey' in Spanish. We like monkeys.
72 Q: When will you ship it?
74 A: We do not know when the code will be shipped. The more
75 contributions we get to the project, the sooner it will ship.
77 A rough estimate is that we might be able to run our C# compiler on
78 Linux by the end of the year. That means running the Windows
79 Executable generated by a Microsoft .NET compiler on the Linux
82 We expect that doing GUI applications will require more work on the
83 class libraries. That could take another six months.
85 Q: How can I contribute?
87 A: Check the <a href="contributing.html">contributing</a> section.
89 Q: You guys should innovate instead of copying.
91 A: In this particular case, we see a clear advantage in the platform
92 and we are interested in using the features of the CLI on open source systems.
94 We have decided that we should spend our limited resources towards
95 implementing an existing specification instead of designing and
98 Designing and implementing our own would be possible, but it doesn't make
99 sense to do that just because the specification comes from a
104 Q: Why is Ximian working on .NET?
106 A: We are interested in providing the best tools for programmers to
107 develop applications for Free Operating Systems.
109 For more information, read the project <a
110 href="rationale.html">rationale</a> page.
112 Q: Will Ximian be able to take on a project of this size?
114 A: Ximian will not be able to taken on the whole project on its own.
115 Mono will be a free software/open source community project, that is
116 the only way we can hope to implement something of this size. You
117 can <a href="contributing.html">contribute</a> to this effort.
119 Q: What pieces will Ximian be working on?
121 A: We will focus on building a development and execution
124 The idea is to get Mono to a state of that would allow
125 third parties to actually be able to use it real-world development.
127 Q: Why does Ximian even care?
129 A: We like the features that the CLI and its related technologies
130 bring to the table. An exciting addition to the developer toolkit.
131 The goal of Mono is to bring this technology to non-Windows
132 platforms (although we hope Mono will also run on Windows, for
133 debugging and comparative purposes).
138 Q: Will I be able to write proprietary applications that run with
141 A: Yes. The licensing scheme is planned to allow proprietary
142 developers to write applications with Mono.
144 Q: What license is Mono on?
146 A: The C# Compiler is released under the terms of the GPL. The class
147 libraries will be under the LGPL or the GPL with a special
148 exception. The runtime libraries are under the LGPL.
150 Since the LGPL is not suitable for embedded systems development, we
151 are also licensing the libraries under the GPL with the following exception:
153 If you link this library against your own program, then you do not
154 need to release the source code for that program. However, any
155 changes that you make to the library itself, or to any native
156 methods upon which the library relies, must be re-distributed in
157 accordance with the terms of the GPL.
159 This is similar in spirit to <a
160 href="http://www.gnu.org/software/classpath/classpath.html">GNU
163 Q: But in Object Oriented Programming I need to subclass your library
164 functions, does that mean that I am making modifications to your
165 library and hence I would have to distribute my sources under the
168 A: No. Object Oriented Programming in the class library is a well
169 understood interface barrier, so you can actually develop
170 proprietary applications with the Mono libraries.
172 Q: Will you accept code under the XXX License?
174 A: If the XXX License is compatible with the license we use in that
175 specific piece of code, then yes. If you want to use the BSD license, make
176 sure you use the BSD license without the advertisement clause (The
177 `Ousterhout License').
181 Q: If applications use Mono, does that mean that I have to pay a service fee?
183 A: No. Mono is not related to Microsoft's initiative of
184 software-as-a-service.
186 Q: If you implement .NET, will I depend on Microsoft Passport to run my software?
188 A: No. The .NET Framework is a runtime infrastructure and collection
189 of class libraries. Passport may be required to access certain web
190 services written for that framework, but only if the programmer
191 chooses Passport as the authentication mechanism.
193 Q: Is .NET just a bunch of marketing slogans?
195 A: Although the `.NET initiative' is still quite nebulous, The .NET Framework
196 has been available for some time. Mono is not an implementation of the .NET
197 initiative, just the development framework.
199 Q: What is a 100% .NET application?
201 A: A `100% .NET application' is one that only uses the APIs defined
202 under the System namespace and does not use PInvoke. These
203 applications would in theory run unmodified on Windows, Linux,
204 HP-UX, Solaris, MacOS X and others.
206 Q: But Microsoft will release a port of the real thing under the
207 `Shared Source' license, why bother with anything else?
209 A: The Shared Source implementation will not be usable for commercial
210 purposes. We are working towards an implementation that will grant
211 a number of rights to recipients: use for any purpose,
212 redistribution, modification, and redistribution of modifications.
214 This is what we call <a
215 href="http://www.gnu.org/philosophy/free-sw.html">Free Software</a>
219 Q: Is this a free implementation of Passport?
221 A: No. Passport is part of Microsoft's Hailstorm initiative. Mono
222 is just a runtime, a compiler and a set of class libraries.
224 Q: Will the System.Web.Security.PassportIdentity class, mean
225 that my software will depend on Passport?
227 A: No. That just means that applications might use that API to
228 contact a Passport site.
230 As long as your application does not use Passport, you will not
233 It might even be possible to implement that class with
234 a set of dummy functions, or use an alternate Passport implementation.
236 We do not know at this time whether the Passport protocol is
237 documented and whether we will be able to talk to
240 Q: What is your opinion?
242 A: You can read my personal <a href="passport.html">opinion on
245 Q: Will Mono running on Linux make Passport available for Linux?
247 A: The Passport toolkit for Linux-based web servers is available from
250 Again, Mono has nothing to do with Passport.
254 Q: Will Mono allow me to run Microsoft Office on Linux?
256 A: No, it will not. Microsoft Office is a Windows application. To
257 run Windows applications on Intel Unix systems refer to <a
258 href="http://www.winehq.com">the Wine Project</a>
262 Q: How is this related to GNOME?
264 A: In a number of ways:
266 * Mono will use existing
267 components that have been developed for GNOME when it makes
268 sense. For example on X systems, we will use Gtk+ and
269 Libart to implement Winforms and the Drawing2D API.
271 For database access, we will use LibGDA (not really
272 depending on GNOME, but related to).
274 * This project was born out of the need of providing improved
275 tools for the GNOME community.
277 * We would like to add support to our CLR implementation to
278 deal with GObjects (in GNOME 1.x, they are called
279 GtkObjects), and allow Mono developers to provide GObjects
280 or use and extend existing GObjects.
282 Q: Has the GNOME Foundation or the GNOME team adopted Mono?
284 A: Mono is too new to be adopted by those groups. We hope that the
285 tools that we will provide will be adopted by free software
286 programmers including the GNOME Foundation members and the GNOME
289 Q: Should GNOME programmers switch over to Mono?
291 A: Mono will not be ready even within the next six months, and a
292 complete implementation is probably one year away.
294 We encourage GNOME developers to continue using the existing tools,
295 libraries and components. Improvements made to GNOME will have an
296 impact on Mono, as they will provide the "backend" for various
299 Q: Will Mono include compatibility with Bonobo components?
301 A: Yes, we will provide a set of classes for implementing and using
302 Bonobo components from within Mono.
304 Q: Does Mono replace Bonobo?
306 A: Bonobo is very focused on cross-application component reuse. Mono
307 will provide a Bonobo framework to allow you to develop Bonobo
308 components and use Bonobo components on Unix.
310 Mono should allow you to write Bonobo components more easily, just
311 like .NET on Windows allows you to export .NET components to COM.
315 Q: Is Mono a way of running Java applets?
321 Q: Is Mono just a new way of writing Web Services?
325 Q: If this implements the SDK classes, will I be able to write and
326 execute .NET Web Services with this?
330 When the project is finished, you will be able to use the same
331 technologies that are available through the .NET Framework SDK on
332 Windows to write Web Services.
336 A: Soup is a library for GNOME applications to create SOAP server and
337 SOAP clients. You can browse the source code for soup using <a
338 href="http://cvs.gnome.org/bonsai">GNOME's Bonsai</a>
342 A: Yes. The CLI contains enough information about a class that
343 exposing it to other RPC systems (like CORBA) is really simple, and
344 does not even require support from an object.
346 We will be implementing CORBA interoperation as an extension to the
347 Mono classes so that we can integrate with Bonobo, just like
348 Microsoft provides COM interoperation classes and support
351 Q: Can I serialize my objects to other things other than XML?
353 A: Yes, although the serializing tools have not yet been planned, and
354 you would probably have to implement them yourself.
358 Q: Will it be possible to use the CLI features without using bytecodes
361 A: Yes. The CLI engine will be made available as a shared library.
362 The garbage collection engine, the threading abstraction, the
363 object system, the dynamic type code system and the JIT will be
364 available for C developers to integreate with their applications if
367 Q: Will you have new development tools?
369 A: Hopefully Free Software enthusiasts will contribute tools to
370 improve the developer environment. These tools could be developed
371 initially using Microsoft implementation of the CLI and then
372 executed later with Mono.
376 Q: What about using Java? After all there are many languages that
379 A: You can get very good tools for doing Java development on free
380 systems right now. <a href="http://www.redhat.com">Red Hat</a> has
381 contributed a <a href="http://gcc.gnu.org">GCC</a> <a
382 href="http://gcc.gnu.org/java">frontend for Java</a> that can take
383 Java sources or Java byte codes and generate native executables; <a
384 href="http://www.transvirtual.com">Transvirtual</a> has implemented
385 <a href="http://www.kaffe.org">Kaffe</a> a JIT engine for Java;
386 Intel also has a Java VM called <a
387 href="http://www.intel.com/research/mrl/orp">ORP</a>.
389 The JVM is not designed to be a general purpose virtual machine.
390 The Common Intermediate Language (CIL), on the other hand, is
391 designed to be a target for a
392 wide variety of programming languages, and has a set of rules
393 designed to be optimal for JITers.
396 Q: What kind of rules make the Common Intermediate Language useful for
399 A: The main rule is that the stack in the CLI is not a general purpose
400 stack. You are not allowed to use it for other purposes than
401 computing values and passing arguments to functions or return
404 At any given call or return instruction, the types on the stack
405 have to be the same independently of the flow of execution of your
408 Q: I heard that the CIL is ideal for JITing and not efficient for
409 interpreters, is this the case?
411 A: The CIL is better suited to be JITed than JVM byte codes, but you
412 can interpret them as trivially as you can interpret JVM byte
415 Q: Could Java target the CLI?
417 A: Yes, Java could target the CLI. We have details on a <a
418 href="ideas.html#guavac">project</a> that someone could take on to
423 Q: Would you allow other classes other than those in the
426 A: Yes. The Microsoft class collection is very big, but it is by no
427 means complete. It would be nice to have a port of `Camel' (the
428 Mail API used by Evolution inspired by Java Mail) for Mono
431 You might also want to look into implementing CORBA for Mono. Not
432 only because it would be useful, but because it sounds like a fun
433 thing to do, given the fact that the CLI is such a type rich
436 For more information on extending Mono, see our <a
437 href="ideas.html">ideas</a> page.
439 ** Mono and portability
441 Q: Will Mono only work on Linux?
443 A: Currently, we are doing our work on Linux-based systems and
444 Windows. We do not expect many Linux-isms in the code, so it
445 should be easy to port Mono to other UNIX variants.
447 Q: What about Mono on non X-based systems?
449 A: Our main intention at Ximian is to be able to develop GNOME
450 applications with Mono, but if you are interested in providing a
451 port of the Winform classes to other platforms (frame buffer or
452 MacOS X for example), we would gladly integrate them, as long
453 they are under a Free Software License.
455 ** Reusing existing Code
457 Q: What projects will you reuse or build upon?
459 A: We want to get Mono in the hands of programmers soon. We are
460 interested in reusing existing open source software.
462 Q: What about Intel's research JIT framework, ORP?
464 A: At this time, we are investigating whether we can use elements of
465 ORP for Mono. ORP is a research JIT engine that has a clear
466 defined API that splits the JIT from the GC system and the actual
467 byte code implementation. It is a research product.
469 Q: What about using GNU Lightning?
471 A: We are also researching <a
472 href="http://www.gnu.org/software/lightning/lightning.html">GNU
475 <a name="ximian-and-microsoft">
476 ** Ximian and Microsoft
478 Q: I read that Microsoft is helping out Ximian, is this true?
480 A: Initial contact between David Stutz and Miguel de Icaza happened.
481 It was a friendly conversation. Microsoft is interested in other
482 implementing .NET and are willing to help make the ECMA spec more
483 accurate for this purpose.
485 We were initially contacted by Sam Ruby at the ECMA TG3 committee
486 to discuss the same issue. And we are glad to have good contacts
487 to ask questions about the specs.
489 Q: Do you fear that Microsoft will change the spec and render Mono useless?
491 A: No. Microsoft proved with the CLI and the C# language that it was
492 possible to create a powerful foundation for many languages to
493 interoperate. We will always have that.
495 Even if changes happened in the platform which were undocumented
496 (which is very unlikely), the existing platform has a value on its
499 Miguel once explained its motivation for working on Mono to Dave
500 Winer, and his mail got posted <a
501 href="http://scriptingnews.userland.com/stories/storyReader$1275">here</a>
503 Q: Didn't Miguel de Icaza say that `Unix Sucks'?
505 A: Yes, he did, as a catch phrase in his opening remark on the Ottawa
506 Linux Symposium. His talk focused on various ways to improve Unix.
508 There is a paper describing some ways to improve Unix at:
510 <a href="http://primates.ximian.com/~miguel/bongo-bong.html">
511 http://primates.ximian.com/~miguel/bongo-bong.html</a>
513 Q: Didn't Ximian's Miguel work for Microsoft?
515 A: Actually, Nat Friedman (Ximian's co-founder) did work as an
516 intern for Microsoft for one summer but Miguel did not.
518 Q: Did Nat and Miguel meet at Microsoft?
520 A: They met online on the Linux IRC network; They met in person for
521 the first time in 1997.
523 ** Mono and Microsoft
525 Q: How can you expect Mono to compete with Microsoft, wont this
526 require an effort too large?
528 A: You are right. Mono will never become a reality without the help
529 of other contributors. Ximian is a small company that can not
530 finish Mono alone. We will be working with members of the
531 community to deliver the product.
533 Q: Is Microsoft and Corel involved in the Mono implementation?
537 Q: Are you writing Mono from the ECMA specs?
539 A: yes, we are writing them from the ECMA specs and the published
540 materials in print about .NET
544 Q: What is the difference between CLR (Common Language Runtime) and
545 CLI (Common Language Infrastructure)?
547 A: CLI is the specification of an execution system. The Microsoft
548 implementation of this specification is named CLR.
550 Unless we come up with our own acronym, we could just call ours
551 also CLR, just because it would do exactly the same thing the
552 Microsoft implementation does.
557 Q: Should someone work on a GCC front-end to C#?
559 A: I would love if someone does, and we would love to help anyone that
560 takes on that task, but we do not have the time or expertise to
561 build a C# compiler with the GCC engine. I find it a lot more fun
562 personally to work on C# on a C# compiler, which has an intrinsic
565 Q: Should someone make a GCC backend that will generate CIL images?
567 A: I would love to see a backend to GCC that generates CIL images. It
568 would provide a ton of free compilers that would generate CIL
569 code. This is something that people would want to look into
570 anyways for Windows interoperation in the future.
572 Q: What about making a front-end to GCC that takes CIL images and
573 generates native code?
575 A: I would love to see this, specially since GCC supports this same
576 feature for Java Byte Codes. You could use the metadata library
577 from Mono to read the byte codes (ie, this would be your
578 "front-end") and generate the trees that get passed to the
581 Ideally our implementation of the CLI will be available as a shared
582 library that could be linked with your application as its runtime
585 Q: But would this work around the GPL in the GCC compiler and allow
586 people to work on non-free front-ends?
588 A: People can already do this by targeting the JVM byte codes (there
589 are about 130 compilers for various languages that target the JVM).
591 Q: Why are you writing a JIT engine instead of a front-end to GCC?
593 A: The JIT engine and runtime engine will be able to execute CIL
594 executables generated on Windows.
596 ** Mono and Portability
598 Q: Will Mono work on other variants of Unix?
600 A: Yes. We do not expect to add any gratuitous incompatibilities.
602 Q: Will Mono run on Windows?
604 A: Hopefully yes. Currently some parts of Mono only run on Windows
605 (the C# compiler is a .NET executable) and other parts have only
606 been compiled on Linux, but work on Windows with Cygwin.
608 Q: Will Mono depend on GNOME?
610 A: It will depend only if you are using a particular assembly (for
611 example, for doing GUI applications). If you are just interested
612 in Mono for implementing a `Hello World Enterprise P2P Web
613 Service', you will not need any GNOME component.
617 Q: How fast will be Mono?
619 A: We can not predict the future, but a conservative estimate is that
620 it would be at least `as fast as other JIT engines'.
622 Now, wishfully thinking I hope that we will ship various JITs with
623 Mono just like Microsoft has done. A fast JITer when maximum
624 performance is not needed, but fast load times are important; And
625 an optimizing JITer that would be slower at generating code but
626 produce more optimal output.
628 The CIL has some advantages over the Java byte code: it is really
629 an intermediate representation and there are a number of
630 restrictions on how you can emit CIL code that simplify creating
633 For example, on the CIL the stack is not really an abstraction
634 available for the code generator to use at will: it is just a way
635 of creating a postfix representation of the parsed tree. At any
636 given call point or return point, the contents of the stack are
637 expected to contain the same object types independently of how the
638 instructions was reached.
640 ** Assorted questions
642 Q: You say that the CLI allows multiple languages to execute on the
643 same environment. Isn't this the purpose of CORBA?
645 A: The key difference between CORBA (and COM) and the CLI is that the
646 CLI allows "data-level interoperability" because every
647 language/component uses the same data layout and memory management.
649 This means you can operate directly upon the datatypes that someone
650 else provides, without having to go via their interfaces. It also
651 means you don't have to "marshall" (convert) parameters (data
652 layouts are the same, so you can just pass components directly) and
653 you don't have to worry about memory managment, because all
654 languages/components share the same garbage collector and address
655 space. This means much less copying and no need for reference