Slashdot Mirror
Jepson Rebuts Petreley On The Dangers Of Mono
Tim O'Reilly writes "Brian Jepson, a long-time perl hacker who's been working with Microsoft on some of O'Reilly's books about .Net, and with Miguel on mono, rebuts Petreley's warnings about mono."
← Back to Stories (view on slashdot.org)
What I don't get is why the HELL they are implementing the CLR. There's already a GNU endorsed, open source VM architecture - Kaffe. GCC 3.0 does java bytecode's now, too. It's a massive wheel-reinvention to go implement an MS technology.
KDE and GNUStep are both better than anything the NGOME camp's put out anyway.
Go back and reread the article. You obviously don't understand the technical details.
The Common Language Runtime and web services and so on DO NOT NEED PASSPORT!
Passport is simply a web service that Microsoft is offering which provides a single-signon. You don't have to use it, it's NOT required. You can use it if you want to.
It is amazing how eagerly people are willing to believe conspiracy theories.
It is not the fault of any of you to do so, but it would be Petreley's duty to do a little research before writing from an uninformed point of view.
I remember I was once sitting in a panel with another famous pundit. The panel was on `Open Source' and they had brought the `experts' on Open Source to talk about it.
Such pundit said a minute before the panel begun to us `I do not see the difference between Peer to Perr and Open Source, to me they are the same thing'.
In Market speak .NET is a Microsoft-wide company initiative. To humans this means that a vision has been set at Microsoft, and every bit of the company is moving towards making that goal happen.
Now, this .NET "vision" encompasses many different areas, let me list some of them for you:
People are confused between two elements: The development platform and Passport. People who claim that Mono is related to passport are making statements that are similar to `Excel is Turbo Pascal'.
The Mono Project is a project that Ximian has launched and is devoting resources to bring the benefits of a new, fresh and powerful development platform to Unix. Why we are doing this? Because we need those tools to bring you the next generation end-user applications. We could do this with C, C++ or Python, but it would take us longer, and we would waste our engineering time.
Passport is a completely different beast. Many people are confused, and asked me `How is Mono related to Passport'? The answer is: it is not related in any way. I have written a document that highlights the problems on Passport, you can get it here: http://www.go-mono.com/passport.html. This has nothing to do with Mono, but people are still confused.
Even worse, some people claim that `Mono will bring Passport to Linux'. Those people have not been paying attention. Passport is already available to be ran on Linux servers. Just go to http://www.passport.com, and download the Linux toolkit.
This bears repetition: Mono is not related to passport.
If you want to have an informed opinion on Mono, please read our FAQ: http://www.go-mono.com/faq.html
miguel.
.NET and Passport are just functions and noone is required by law (yet) to use them.. Example? Ebay now uses them, if they require it then I stop using ebay and I tell them why. If you contact an E-commerce site and tell them they you cant buy their products because of a dimwitted decision to lock out people that dont use MS software they will take notice. (In e-commerce today, one lost sale is felt, and general managers will hang sales people for losing sales.... A GM will probably fire the entire web team if their migration to .net and passport hurts sales even just a little nowdays.
Ecommerce isnt driven by Microsoft and their drivel, It's driven by moola, dollars, that green stuff in your wallet. That is what Ecommerce believes in fully, and that is all they care about.
Ecommerce sites will bail on microsoft instantly if it affects cash flow or the bottom line.
Do not look at laser with remaining good eye.
I swear, it's cliche come true: Those who forget history are doomed to repeat it. It's like the Ximian guys are abused spouses. They're making excuses, thinking they can change the other person, pretending it'll all work out good this time, forgiving this one last time, so on and so on. People get therapy for this, you know. The solution is not to give them one more chance. The solution is to get the hell out and start fresh. We cannot let this codependant wool-gathering to continue where Microsoft is concerned. They cannot be beaten at their game. We need to get them to play our game. There are many things OSS can do better than a monolithic organization. we'll win by focusing on that. Proaction, not reaction.
Hmmm. Time for a new analogy.
People need to remember that MS is out to do one thing and one thing only: Create and maintain shareholder value. That's the only thing publicly traded companies aim to do and the only thing that matters to them. Everything besides that is secondary (contrary to the "We like to make innovative software for the betterment of the world" mantra the PR types like to repeat to us). And the cash goal is just fine; it's what our country was built on and it's created a lot of happy people and started a lot of successful companies. It works pretty well. Until you throw in some incredibly lucrative technology backed by technologically savvy yet altruistic individuals. It's chum for the MS corporate shark.
MS could no more resist the urge to co-opt and/or subvert Mono or related techologies than a Great White could resist a tasty meal at the Ft. Lauderdale Hemophiliac Seniors Beach club. MS will come out with .NET and Passport, and Ximian will have Mono. As soon as organizations start favoring Mono over MS's proprietary tech, MS will devour Mono by any means necessary. They have to, in order to maintiain market share. They have to keep up with the slippery slope, and nobody can tell me they won't do whatever it takes when it comes right down to it. They have to stay moving and eating to survive, and will react badly to anything that threatens their food supply of cash and mindshare. What we are dealing with here is a perfect engine... an eating machine. It's really a miracle of business evolution. All this machine does is extend and embrace and extinguish other competeing technologies and that's all. (Apologies to Mr. Hooper.)
Anyway, people would do themselves well to remember the past.
-B
Ash and Hickory, straight-grained and true, make excellent bludgeons, dandy for the cudgeling of vegetarians.
Miguel de Icaza and Nat Friedman response is here
Go ahead, hit me with the -1, Flamebait. I can take it.
.Net. The Linux crowd worked for years just getting Java ported, and it already ran on a flavor of Unix)
In all seriousness, though, I cannot even remotely imagine Mono actually working for anything more complicated than Hello World apps. Consider that our experience with powerful cross-platform languages so far basically starts and stops with Java. Before anybody starts spouting out examples of their favorite cross-platform programming language: unless you've written a word processor or equally heinous client-side app and had it work on other systems on the first or second try, it does not count.
So let's consider Mono in light of our current experience with Java (disclaimer: I am a professional Java developer). Java's basic operations were pretty simple to get right: right out of the gate most non-GUI apps were likely to work on all platforms (as long as they weren't too big...). The AWT, however, was a disaster. GUIs are a lot more complicated than most non-GUI programmers realize, and tiny inconsistencies between platforms can easily break things. The AWT was the simplest API imaginable, and after two years of Sun's (and many other's) work on it getting GUI apps to work was still very hit-and-miss. Keep in mind that this API was so simple that it didn't even support clipboards, drag-and-drop, or even PRINTING in its first release. It didn't have tables, or real file dialogs, or trees, or split panes, or a million other things we take for granted. And it still didn't work.
Only after the incredible investment of time that Swing took was Java finally able to *really* get things right between platforms. For the non-Java-nuts among us, Swing is a component library written completely in Java. The native system libraries are used only for very basic painting and input operations. In other words, getting native GUIs to work the same way on multiple platforms was essentially impossible for Sun: they gave up and wrote it all from scratch in Java.
The situation with Mono is even worse, because the API is by Microsoft. I confess that I am not familiar with WinForms; if it's an incredibly elegant, simple API please correct me. However, most GUI libraries are REALLY, REALLY complicated; again, if you've never tried writing something like a cross-platform word processor, you just don't understand how tough it can be. The idea of everything working exactly the same between Mono and Windows (when Winforms simply delegates to Window's native API) is almost laughable.
I'll close with a simple example of what the Java team faced and why Swing even exists. The AWT TextArea naturally just used the native Windows, Macintosh, or whatever text area to get its work done. Exactly, I believe, as WinForms does. Problem: the Windows implementation of TextArea wouldn't work with more than 32K of text under Win95 and Win98. Unix programmers would happily dump megabytes of text into their (perfectly working) text areas, which would then blow up on Windows. They would then bitch at Java for not working cross-platform. This particular issue with Windows may since have been resolved, but in its time it was a very large problem. There were many, many other similar things -- which is why for the most part nobody uses native GUIs in Java anymore.
No disrepect to the Mono team intended, but I'm inclined to believe that any complicated apps won't have a prayer in hell of working. Again, the non-GUI stuff may not be a problem, but Unix doesn't need any help in that department. What Unix needs is good client apps, and I think the Mono team is completely out of touch with reality if they think it's as simple as hacking together a Winforms implementation overnight.
(And please don't bother responding with anything along the lines of "Yeah, but it's open-source so it must be better than what Sun can do!" Bullshit. Sun, Netscape (now AOL) IBM, the Apache team, and a million other groups are all working on Java, and there's no way such incredible development resources will ever be pooled together for a simple port of
--- egomaniac
ZFS: because love is never having to say fsck
Is it just me, or does anyone else have vivid flashbacks of the Outbreak monkey every time you see the Ximian logo next to a headline containing "Mono"?
Remember how funny that was? Remember how funny it was that the Net portrayed in that movie was so fast you could practically do full-motion video? That was a real scream when you got home to log onto this net with your 56K. And now that you've got a cable modem, well, things are looking a bit faster. And now that you have a 1.2 GHz processor and 512 Mb of RAM the scene where Sandra opens about 50 windows whilst seeking out BillG, er, the bad guy's identity doesn't look quite as foolish any more.
And the idea of using one application to hack all those different real-world apps -- police, jails, drivers' licenses, hospital records, all stored in a completely compatible and accessible form. Baloney, you remember thinking, thinking of all those old WordPerfect 5.1 docs the boss wants you to convert to Word for Windows 6.0.
Well, it's not so funny any more because that's exactly what .NET is. MS want not only to perfectly integrate all your apps, they want to hold the data on their servers. It's almost as if BillG saw that movie and thought, "gee, what a swell idea."
I've lost count of the number of apps I've run across written with Access, or SQL, when flat files would have been more efficient and made more sense. But flat files tend to be proprietary in form, distribution, and content. If you've ever gone into an 80's-era accounting system you know what I'm talking about -- hundreds of files with enlightening names like X878190.DAT.
But put it all on a SQL server, especially one owned by BillG, and things become much clearer. Every record is tagged as to type and you can much more quickly work out what data go where. It wouldn't be perfect, but tricks like switching the photos on a couple of driver's licenses would become much simpler (assuming, of course, that you go for the last piece of the puzzle and install your back door).
Microsoft has the ultimate history of irresponsibility. Since their first days in business they've specialized in producing slow, buggy, unfinished code which they've rammed down our throats through brilliant (and often illegal) marketing tactics. Now they want to centralize all the data in the world under their own aegis. Excuse me while I go puke.
Brackets contain world's first nanosig, highly magnified:[.]
Because they said so in a press release. This wasn't about .NET, it was about one of their other efforts -- Hailstorm IIRC.
But like any service, it is something you subscribe to and decide to participate with. Microsoft's hope is that you'll find it conveninient and want to use it. If not, then well what are you worried about?
Well, it's like this. I have credit cards. Now there is no reason on Earth I need credit cards; I'm debt-free and have plenty of money in checking. I have credit cards because of the absurdly large number of things you can't do without them. Like rent a car, reserve a hotel room, or get stuff delivered next-day when you order it online. Microsoft is positioning itself to be the next "credit card," only moreso because they will invade every area of life that involves a computer. Do I get to "opt out" of the police, hospitals, fire stations, ambulance services, ADT (who write paychecks for my employer), or my employer placing itself in Microsoft's grand scheme of things? Of course not. So as in the movie, if BillG really were to put a backdoor into the architecture, he could create a great deal of mischief for me despite my aversion to his wares.
They certainly have a history of producing high quality software. I don't know what software you've been using that you consider low quality.
Well, it's like this.
1970's
1980's
- DOS 1.0, basically stolen
- DOS 2.0, nearly equal to CP/M
- DOS 3.0, broke half the apps that ran on 2.0
- DOS 4.0, broke 3/4 of the apps that ran on 3.0 AND was a memory hog AND full of bugs (but then, "DOS isn't done until LOTUS won't run")
- Their first series of compilers, which were unaware of the 8088 failure to complete string copy operations after an interrupt, and so crashed randomly
1990'sIt's true that they have done a few things right, especially w/r/t Microsoft Office. But I could have told you in 1978 that their corporate style wasn't capable of developing a stable operating system. Early versions of Word tended to blow up a lot too, but then they didn't tend to take the whole universe with them when they did.
You then claim that this software is rammed down users throats. Are you just gullible? I've never had software rammed down my throat.
You've never received a document as a .DOC after you had multiply and explicitly demanded .RTF because of the totally unnecessary virus threat caused by their irresponsible security defaults, you've never had to install a Windoze partition on a machine just to run some app that is unavailable in non-Windoze version which turns out to be indespensible either to you or a coworker, you've never...
Wait a minute. Of course you haven't been forced to because you're so ignorant you actually like this shit. Never mind.
Brackets contain world's first nanosig, highly magnified:[.]
Monkey see monkey do... =) Why should people use a bad copy of a bad software ?