LWN.net Logo

Advertisement

E-Commerce & credit card processing - the Open Source way!

Advertise here

de Icaza: Mono directions

Miguel de Icaza notes the Mono 1.1.10 release with a lengthy document on where the Mono project (at least, the part of it housed at Novell) plans to go from here.
(Log in to post comments)

novell throwing away money on this

Posted Nov 18, 2005 19:01 UTC (Fri) by b7j0c (subscriber, #27559) [Link]

with all of the layoffs there, you would think management would be able to rationalize the real effectiveness of further investment in mono.

.net is not going to be the de facto application delivery platform for the future, even microsoft concedes this, particularly given their mae culpa as of late on webservices (read: google) and the push to drive products like msn search and windows live. even microsoft grudginly accepts now that the browser stack is the delivery platform of the future.

so why bother at this point throwing more money at .net? if i were at novell, i would get mono to a decent level of .net support, then put it in to deep freeze/bug fix mode until someone can demonstrate that anyone is going to derive revenue from this or even care.

novell throwing away money on this

Posted Nov 18, 2005 20:08 UTC (Fri) by migueldeicaza (guest, #33951) [Link]

There are a number of reasons that you might have missed.

Mono is used at Novell to develop various products, and increase our effectiveness:

* ZenWorks for Linux (we get to share the client code base with Windows).
* iFolder (On Mac and Linux, we use Mono; On Windows .NET)
* Our new desktop apps: Banshee, F-Spot and Beagle.

In addition to Mono being a framework that we use at Novell for several projects, it helps ISVs that have products built for Windows-only systems to port their applications to Linux.

This benefits both the Linux users and the Windows vendors.

novell throwing away money on this

Posted Nov 19, 2005 7:57 UTC (Sat) by drag (subscriber, #31333) [Link]

Why snub a new programming language and framework?

Linux has always supported all these different programming languages and whatnot. We have ruby, python, c, c++, perl, fortran, and a whole host of other languages that people use for various purposes.

Plus for Linux-on-the-desktop we need a smart interpreted language for developing applications. I use F-spot, and on a daily basis I use Tomboy. both are very nice and fairly fast. I use Beagle on my desktop, but unfortunatly it's not so hot on my PPC laptop yet.

(of course if you were talking about java no Java app will ever work in my PPC laptop since Java is non-portable. (the code is, but what is needed to run the code is not))

With Gnome before it was a principally a C-only arena.. now they are happily embracing Python and C#.. This allows developers much more time on developing GUIs, creating secure applications and the like rather then chasing around C problems constantly... Plus then you still can use C at the core for things that require good speed.

A example of a python application that I like to use is Straw, the desktop rss application for gnome.

If your a programmer for a business you can now do 'rapid application developement' using C# or Python for a Linux enviroment and tie it into databases and web apps and all sorts of stuff like that that would of been very difficult to do just a few years ago, while on Windows that was commonplace.

Plus with C# it's something that people actually KNOW. Tons and tons of people are taking classes in it. .NET books are commonplace and the number of Windows webdesigners familar with .NET and C# outnumber the number of programmers that are familar with Gnome/KDE and Linux C/C++-isms by quite a large margin. A person familar with .NET/C# in Windows can now go to Linux and make a good application or contribute to a project with almost no prior Linux experiance.

This sort of thing helps make it much easier for people to migrate from Windows business desktop to a Linux business desktop, especially for people with large numbers of users and have a need for lots of custom applications. Lots of people know how to do desktop development for Windows, almost nobody knows how to do desktop development for Linux. Now the bar for entry is much lower then it used to be.

You know.. migrate people off of Windows/Microsoft infrastructure to a Suse/Novell based enviroment is the whole goal of being Novell at this moment.

Free Software on PPC (and other archs)

Posted Nov 19, 2005 9:35 UTC (Sat) by mjw (subscriber, #16740) [Link]

(of course if you were talking about java no Java app will ever work in my PPC laptop since Java is non-portable. (the code is, but what is needed to run the code is not))
See Cacao to launch Eclipse 3.1 on PowerPC and ObjectWeb conf presentation on gcj, major application servers now run on the various GNU/Linux platforms for the various architectures (including ppc) supported by gcc.

novell throwing away money on this

Posted Nov 21, 2005 3:52 UTC (Mon) by khim (subscriber, #9252) [Link]

(of course if you were talking about java no Java app will ever work in my PPC laptop since Java is non-portable. (the code is, but what is needed to run the code is not))

What's wrong with PPC ? I had Java working there in Gentoo just fine... IBM's JDK...

novell throwing away money on this

Posted Nov 18, 2005 23:56 UTC (Fri) by adamoell (guest, #33953) [Link]

I don't think you can say that Microsoft has given up on .NET as an application platform for the future. Your example of web services is apposite: how do you develop web services on the Microsoft platform? Answer: .NET. Browser delivery may be the way to go, but that's always been a core part of the .NET strategy anyway.

However, I don't really think that many folk at MS believe that the browser is the answer to all clients. Look, for instance, at Office 12, Smart Clients and many other recent, strategic technologies from MS - clearly the browser is seen as a crucial part of the picture, but not the only piece.

Lastly, I think many people who work primarily in non-Microsoft environments would be absolutely *astonished* to see the uptake that the .NET platform has had in its relatively short lifespan. Just search a job site for .NET developers - there is now nearly as much demand as for Java developers, in spite of the fact that Java is much, much older.

My view is that .NET is clearly an important platform which is going to be around for a long time; many organisations (of all sizes) have strategic commitments to it; it needs to be in the Free Software "toolkit" and the Mono project has done absolutely brilliant work on this. Perhaps Novell may struggle to derive revenue from it, but there are certainly plenty of us who care!

novell throwing away money on this

Posted Nov 19, 2005 22:20 UTC (Sat) by b7j0c (subscriber, #27559) [Link]

>> Your example of web services is apposite: how do you develop web services on the Microsoft platform? Answer: .NET.

the point is no one cares how these services get delivered. do you choose to use hotmail because it is being served from a .net platform?? no, if it works in your browser, cheers. firefox has pushed microsoft into accepting (grudgingly) browser portability as well.

linux already has a platform for webservices delivery that is more well known and more widespread than .net - LAMP.

>> I don't really think that many folk at MS believe that the browser is the answer to all clients. Look, for instance, at Office 12

yes but this is a dead market, microsoft can barely motivate its own userbase to upgrade.

novell throwing away money on this

Posted Nov 19, 2005 2:23 UTC (Sat) by larryr (guest, #4030) [Link]

.net is not going to be the de facto application delivery platform for the future, even microsoft concedes this [...] microsoft grudginly accepts now that the browser stack is the delivery platform of the future.

What the hell is a "delivery platform"? dotNET, like Java, is software libraries and a runtime environment for applications, independent of the mechanism through which they interact with a user. I think between the two of them they will be the foundation for the majority of application software developed over the next 10 years, especially software which is exposed to the user over the web and through software running in the web browser-- software which may itself have been generated by Java and dotNET software (as "AJAX"). It seems to me that both Java and dotNET are primarily designed to be used in a context where the consumer is using a "browser stack".

Larry

Designed?

Posted Nov 19, 2005 7:29 UTC (Sat) by ncm (subscriber, #165) [Link]

Both Java and Dotnet, as currently deployed, were designed for precisely one thing: to lock development shops into their respective environments. The core language, in each case, has little to do with this; each is a relentlessly pedestrian language just barely strong enough for the job. The environments around them reveal the strategy: they are meant to create makework for five to seven barely-skilled people to do the job one skilled person could do better and faster.

No Java or Dotnet shop can consider using a better tool. What would they do with all the monkeys? What corporate development manager would cut his or her own overall budget by 80%? Most of those coding either language, now, expect to be doing the same until they retire, or are "promoted" into management, just like the COBOLers they are replacing.

A language actually designed for efficient web application development and deployment would look very different from either.

Designed?

Posted Nov 19, 2005 16:40 UTC (Sat) by phgrenet (subscriber, #5979) [Link]

I want to react to your statement that Java and .NET where designed "to lock development shops into their respective environments". This is not exactly correct.

Java was Sun's strategy to stop a growing number of developers from moving from Unix to Windows. This was happening because Windows was (is) the dominant platform, and application authors want to write code for the most popular platform for simple economic reasons. Sun attracted people to Java because Java runs on Unix and Windows (so you don't have to make a choice) and because Java is a higher level language that either makes programmers more productive for most applications, or lowers the minimal required programmer skillset. Well, at least before J2EE.

Dot Net was Microsoft's strategy to prevent a growing numbers of developers from switching from C++ or VB to Java because of this productivity factor. They just developped a clone of Java that offers very little advantages over Java, and told their massive PR machine to convice people that it was much better than Java.

Designed?

Posted Nov 20, 2005 3:51 UTC (Sun) by bojan (subscriber, #14302) [Link]

> Well, at least before J2EE.

He, he, so true! I found myself hammering out same old code over and over again, just to get something out of the database with J2EE in the days of EJB 1.1 (even with CMP beans). I could have probably done the same thing with less effort in C, not so mention simpler environments.

I used to be a huge fan of Java and precisely for the reasons you mentioned. Unfortunately, the beast turned ugly after a while. Did I mention it was always slow and memory hungry like no other environment I've seen? But, some people say it's better now . Don't know - haven't done anything with it in some time...

novell throwing away money on this

Posted Nov 19, 2005 22:22 UTC (Sat) by b7j0c (subscriber, #27559) [Link]

>> It seems to me that both Java and dotNET are primarily designed to be used in a context where the consumer is using a "browser stack".

yes, but no one cares as long as they get back the data correctly, securely, and in a timely manner.

which means .net is at war with everything from mod_perl to RoR to JSPs etc etc etc. there are no economies of scale for microsoft to exploit because they are simply pushing out standards-complient bitstreams (http,html,css,js, etc etc etc). this plays to the cheap and free options, not the vendor solutions.

novell throwing away money on this

Posted Nov 21, 2005 4:02 UTC (Mon) by khim (subscriber, #9252) [Link]

yes, but no one cares as long as they get back the data correctly, securely, and in a timely manner.

Actually it's all buzzwords, nohing more. I've seen more in-house programs developed with .Net and mono on few firms I've consulted then I've seen in-house tools developed with Java and "Web platform" combined.

Why ? It's easy: programmers are there, it works, what's not to like ? With "web platform" even trivial things (like shortcuts) are not-so-easy (they work differently in different browsers and you need horrible kludges and even then sometimes something just does not work).

The fact of the matter is: Microsoft is pushing .NET for future development on Windows platform (both desktop and PDAs). So there will be programmers who will know C# and nothing more. This is fact of life. We can accept it or fight it. Novell accepted this, you and RedHat are fighting this. We'll see what strategy will work the best.

novell throwing away money on this

Posted Nov 21, 2005 6:13 UTC (Mon) by b7j0c (subscriber, #27559) [Link]

>> So there will be programmers who will know C# and nothing more. This is fact of life. We can accept it or fight it. Novell accepted this, you and RedHat are fighting this. We'll see what strategy will work the best.

microsoft's days of making the market are long gone. even if one company could make on language so pervasive that you literally could not escape knowing it, its not going to be microsoft at this point. the desktop pc era is over, whether its getting googled by networked services or mobile devices or whatever, microsoft's days of telling the market what to do are gone.

this idea you posit that programmers will only know c# is absurd. how many video games are written in c#? how many MICROSOFT XBOX 360 games are written in c#? how many MICROSOFT OFFICE products are written in c#? its not even the dominant language in microsoft's own product line!

oh by the way how should be do embedded programming in c#? how about OS kernels? remember you are telling me this is going to be the only language.

.net never happened. its just yet-another api/vm combo that arrived too late to destroy java.

Novell hedging their bets

Posted Nov 21, 2005 7:35 UTC (Mon) by man_ls (subscriber, #15091) [Link]

this idea you posit that programmers will only know c# is absurd.
I understand khim thinks that there will be some programmers who only know C#, just like there are programmers now who only know Java (and make a nice living out of it, by the way). Not that everything will be done in C#.

And why not? It's not maybe a choice for Office or games development; but for business applications on Windows, C# should be good enough. And Novell is going after the business market. So as long as C# is a motive for anyone to stay on Windows (now or in 10 years time), Novell is just removing that motive. Sounds reasonable to me.

Novell hedging their bets

Posted Nov 21, 2005 17:26 UTC (Mon) by b7j0c (subscriber, #27559) [Link]

>> And why not? It's not maybe a choice for Office or games development; but for business applications on Windows, C# should be good enough

except the market for office products is stagnant, unless you are telling me that MS Office is going to be rewritten in C#

otherwise, where are all these new office apps? there is LESS diversity in this market than there was five years ago. its a REPLACEMENT PARTS market now and even microsoft can't get users to upgrade.

Novell hedging their bets

Posted Nov 22, 2005 0:34 UTC (Tue) by khim (subscriber, #9252) [Link]

except the market for office products is stagnant, unless you are telling me that MS Office is going to be rewritten in C#

Hmm. Please remind me when MS Office was rewritten in Visual Basic ? C# is filling role Visual Basic filled 10 years ago and Cobol filled before: in-house programs.

otherwise, where are all these new office apps? there is LESS diversity in this market than there was five years ago. its a REPLACEMENT PARTS market now and even microsoft can't get users to upgrade.

They are where they should be: on workplaces in different companies. You'll not see them on shelf. You'll never replace Windows on desktops without them. That's why we need Gambas and C#. And no, it's not REPLACEMENT PARTS market. Custom-made applications are just that: custom-made. Structure of companies are changin, requirements are changin and so programs need support. Forever. This is invisible part of software iceberg and this is where C# is mostly used today.

May be Microsoft had bigger plans for C#, may be not - but in reality "C# is Visual Basic today" and this is enough to support it.

Novell hedging their bets

Posted Nov 22, 2005 5:36 UTC (Tue) by b7j0c (subscriber, #27559) [Link]

>> This is invisible part of software iceberg

although apparently you can see it....????

Novell hedging their bets

Posted Nov 23, 2005 17:22 UTC (Wed) by hummassa (subscriber, #307) [Link]

>> although apparently you can see it....????

Only if you are under the water :-)
It's my case: I work making custom in-house software (Delphi+Oracle). There will /never/ be a shrink-wrap software that fills the requirements of (most of) our systems, so do we have a team of 15 programmers making and maintaining our custom software.

Novell hedging their bets

Posted Nov 22, 2005 20:38 UTC (Tue) by man_ls (subscriber, #15091) [Link]

Please read again:
It's not maybe a choice for Office or games development; but for business applications on Windows, C# should be good enough
Maybe not for office development, definitely yes for business applications. Who cares about the market for office products? It's for in-house business applications where jobs abound. There it's basically C# or Java; and those users are probably game for a Linux variant, if they can stage a smooth transition. Mono should help there.

novell throwing away money on this

Posted Nov 22, 2005 0:40 UTC (Tue) by khim (subscriber, #9252) [Link]

oh by the way how should be do embedded programming in c#? how about OS kernels? remember you are telling me this is going to be the only language.

Gosh. Are you that stupid ? I never said it'll be "the only language".

1000 programmers are writing kernel (hardcore C).

10000 programmers are writing "commodity applications" (office, browsers, etc). C/C++ (not hardcore), Python, JS - you name it.

100000 programmers are writing simple in-house application in C# (10 years ago they used Visual Basic). May be 10% of them are ready to use Python or even C/C++ but most do know only C# and the only reason they are not using Visual Basic is difference between Visual Basic and Visual Fred: Microsoft forced switch from Visual Basic to C# and they'll abandon C# only if they'll be forced to switch again.

Do you seriously imply that 100000 programmers "do not matter one bit" ? Looks like absurd position to me...

novell throwing away money on this

Posted Nov 22, 2005 5:34 UTC (Tue) by b7j0c (subscriber, #27559) [Link]

>> Do you seriously imply that 100000 programmers "do not matter one bit" ?

no, i am implying your fictional statistics and porous logic do not matter one bit.

novell throwing away money on this

Posted Nov 22, 2005 10:04 UTC (Tue) by gravious (guest, #7662) [Link]

hey b7j0c,

>> no, i am implying your fictional statistics and porous logic do not matter one bit.

I'll do better than imply. What are you basing _your_ idiotic logic on? How many .Net or mono projects have you worked on? My guess is none. My guess is based on the logic that you continually talk out of your ass. Go back to slashdot where you belong and stop slagging off khim in that manner. If you want porous logic, look in the mirror sometime. (And before you slag me off - yes, I have used c# and it rocks and no i'm not a m$ fan-boy. I think I'll continue to trust Miguel De Icaza's judgement and code over your vapourware. Thanks for your contributions anyhow.)

best wishes,
Anthony

de Icaza: Mono directions

Posted Nov 21, 2005 13:14 UTC (Mon) by bangert (subscriber, #28342) [Link]

in a related note: we might just get most of mono's advantages (CLR) within our good old gcc
http://www.kdedevelopers.org/node/1628

i wish red hat, apple and SUSE hat teamed up two or three years ago... throw in sun and ibm
and all off us would be better off...

but then, it's all about platform control. always has, always will....

de Icaza: Mono directions

Posted Nov 23, 2005 16:47 UTC (Wed) by nix (subscriber, #2304) [Link]

Executing the intermediate structures that LLVM produces would be... extremely painful. It's not designed for that.

You might as well try to execute GCC TREEs. (Oddly enough, nothing does that either; we have this thing called a compiler to turn them into executable code.)

Copyright © 2005, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds