Microsoft's Mountain man sees Jobsian past in .NET Nearly 10 years ago, Dan'l Lewin was given an almost unenviable role - to be Microsoft's ambassador in Silicon Valley, a place hostile towards his new employer. But this former senior Apple executive asked for it. Lewin was named vice-president of .NET development in January 2001, with the job of bringing start-ups and …
Sorry but as a long-time Apple fan(boy-no-more) I must comment on the historical perspective of a sentence in this article.
You mention:
"as director of education marketing and sales between 1981 and 1984, he ran Apple's University Consortium program, putting Macs into the hands of students and teaching staff. This solidified the Mac's presence in learning."
Nice.. But the mac has been launched in January1984 being hardly available to mass-market for the first three months. Mr. Lewin (which I profoundly respect for his work) might not have been very influential during such a brief time lapse.
At the time the Apple II was very strong in the traditional educational market (up to high schools); the Mac has been promoted upwards, the target being University and research. Nonetheless resistance has been high in the beginning due to the very limited memory the Mac offered. My perception is that it wasn't until the "Fat Mac" was available that the educational market found the platform of any use.
Whatever. .Net is today Microsoft's best weapon. Redmond (or should I say: Mountain View) is and will be relevant because it offers a continuous development model that starts from the desktop, passes through the Cloud and lands to mobile (some work to be done there but eventually they will deliver).
Gabriele
In public
We're the market leader due to our state of the art technology.
In private (at DoJ anti trust hearings)
Our biggest strength was our army of sales staff with well developed relationships with the Finance and Accounts departments.
Microsoft's greatest strength is still its near monopoly on the desktop and creation of a generation of IT types who know *nothing* of any non-Microsoft servers and little about alternate desktops.
Ballmer knows this. He may be know technie but he understands the strenghth of market dominance.
Anyone looking at a cross platform (or at least platform fairly neutral) solution should look at the history of something. In .Net's case what else does it run on.
...says who?
Seems to me things a shaping up a little proprietary to me.
Google (services) on Google (cloud) in Google (browser) on Google (variant of Linux).
Microsoft (services) in Microsoft (browser) on Microsoft (OS) with Microsoft (server stack).
Oracle (middleware) on Oracle (database) on Oracle (OS) on Oracle (hardware)
Redhat (middleware) on Redhat (OS) inside Redhat (Virtualisation stack)
etc. for any large company you can think of.
"Open standards" appear to be very bad for business. The goal that the companies with actual money are working on seems to be:
Make a giant stack of interconnected software and hardware.
Obtain customers.
Lock customers in.
????.
Profit.
Where or how or what exactly is the big whozawhatzit that is magically going to alter this? Seems to me the 00’s were about good as it’s ever going to get for “free (libre) and open.” The 201Xs are all going to be about stupendously massive consolidation of computing into big, integrated mega corps that will only work together when a government forces them too.
I'd honestly be very happy to be wrong about all of this.
You are correct on certain things like, say, everything Google does is not exactly 'open' to put it mildly.
However, if you're going to mention Redhat middleware, OS and virtualisation you should probably mention that all three of these things are completely open and can and are used as individual components either with some other flavor of OS, middleware or virtualisation. This means no lock in.
To make it clear, JBos is open source, completely open. Yes Redhat owns it, you do not have to used Redhat linux to use JBos. The OS is a distribution of linux. Enough said. There's nothing that locks you into Redhat Linux unless you count support contracts that most companies insist on having and the name 'Redhat' just makes them feel better.
Then virtualisation - unless I am mistaken Redhat used to use Xen as it's main virtualisation software. Xen is again open source, owned by Citrix. No lock in. You can startup a xen instance that used to run on redhat on any other linux running the same xen version.
Now I think they use KVM. Again, this is open source and part of the mainline kernel. No lock in.
A bit different from a closed source DB, a closed source OS (solaris), closed source middleware (oracle's), a not really 'open' open browser (chrome) and associated services in the 'cloud' or as it was known all of two years ago, the 'Internet'.
Redhat is "open," (as in FOSS,) but it is not "interoperable."
Oh, it may be open source, and you can certainly fork the project, or tinker with any part of it you wish...but the various strands of Linux have diverged from one another quite a bit. The simple fact that the various distributions are package (and sometimes binary) incompatible makes things difficult unless the project is actively maintained or ported. (Or you are willing to toss gobs of time and money at it to do yourself.)
Redhat-based disties are Redhat based disties. Debian based is Debian based. Even simple things like vi behave differently between them. For all intents and purposes I consider the various distie "strands" as completely separate and non-interoperable operating systems. (Just as I consider Google's strand of Linux to be separate, distinct, and non-interoperable.) (I often compare it to NT-based Windows versus 9.X-based windows.)
Linux doesn't really talk all that well to anything that isn't Linux, (not the fault of the various projects underway to solve this...but the proprietary companies constantly release a new standard, then document it badly.) Lack of interoperability with proprietary platforms (regardless of who is to blame for that) further isolates Linux from everything else, making it just as much of a self-contained silo as the proprietary stacks.
I understand that if I have committed to JBoss, or KVM, I can take my information in those apps and move it to another distro. With enough tweaking, maybe it could be made to work on a SuSE stack or a Debian stack…but this still requires a massive outlay of testing and certification time. Time is money, especially in corporate IT. Do all the dependency programs and libraries exist? Have they been ported to the appropriate platform? Are there behavioural discrepancies in file system or file handling, etc? Even *if* I can get my app running on two separate disties, there is a whole other round of testing to ensure it runs in the same fashion. (For an example, see vi’s behavioural differences on Debian disties versus Redhat disties.)
Linux isn’t free...at the enterprise level, (where things like JBoss and suchlike play,) you not only buy Enterprise licensing, you spend a lot of time and money testing and certifying. For all that effort, I still can’t make JBoss interoperate with anything from MS, Oracle Google etc. Even if I have the pleasure of (maybe) being able to port my app from one Linux distro to another...what interoperability does that actually buy you?
Please understand I am not bashing Linux in any way here. My statements above come from the fact that my job entails a lot of interoperability work. Getting the various strands of Linux to coexist, talk to Solaris, Windows, BSD and $deity only knows what else is a big part of my year. In my mind, open source or not...from any practical standpoint Linux distribution strands are isolated entities.
I suppose so. It's got the vendor lock-in. Pretty much the only take up in the first ten years has been amongst the VB crowd, who already *had* all that vendor lock-in, as they discovered when the vendor in question took away their language and ordered them all to use .NET instead.
Anyone with a choice has used a language that is either standard or free. If you are a software house, betting the farm on someone else's business plans just doesn't make sense. I note that MS themselves conspicuously failed to throw all their eggs into the IA-64 basket.
"Pretty much the only take up in the first ten years has been amongst the VB crowd, who already *had* all that vendor lock-in, as they discovered when the vendor in question took away their language and ordered them all to use .NET instead."
Actually, MS were pretty brave to do that, VB was one of the most popular languages back in 2001 and they risked alienating (and did) a tremendous number of developers who could jump ship. Having developed in VB and VB.Net, I will say that VB.Net is a much more elegant and modern language. VB had become an illogical mess and needed a ground up rewrite. I think as well, there's been plenty of take up of C# and not just from ex VB developers. I'll say this quietly, its not unknown for some Java developers to prefer C# and .Net.
"In .Net's case what else does it run on." ?
Well er let me see...oh that's it, how about Linux!! Check out the Mono project John!
Oh and if that doesn't float your boat, you could always write ASP.Net web apps. They'll run in all the major browsers across all the main platforms.
But having said that, I think MS should push .Net's cross platform capabilities much much further, particularly on mobiles. Forget the iPhone it's a closed book but wouldn't it be great if MS agressively developed .Net to run on all the other major mobile platforms.
.Net runs only on Windows. End of discussion.
Silverlight is slightly better, running on Windows and OS X.
Neither run on Linux, and probably never will.
Despite the hard work done by the Mono and Moonlight teams, they are simply pissing in the wind. They are always at least one version behind and have serious compatibility short-comings that render any attempt to use a .Net/Silverlight application on Linux futile.
This isn't the fault of those teams, it is MS policy. Having these projects around allows MS to crow about being cross-platform without actually having to be cross-platform at all. The documentation is still withheld and there is precious little (if any) support.
When I can go to an MS site, download an MS made .rpm/.deb/whatever and install a current package that has 100% compatibility with the same version on Windows THEN we can say that .Net runs on non-Windows OSs.
Until then the best we can say is "A sub-set of a previous version that had to be reverse engineered, may run on a non-Window platform but compatibility cannot be guaranteed and the execution may suffer instability, errors and crashes because of this. Therefore the deployment must be considered a technical exercise and must not be used in production in any capacity."
Oh, and before people mentions in, WINE should not be used in production either. And that's according to WINE themselves!
If, of course, you develop in Mono or Moonlight, THAT should work cross-platform. But then you are not developing a .Net solution, it's a Mono solution.
Sometimes el reg comes out and suddenly, inexplicably fails the snark. Instead of biting, it softly pecks, cooing. Oh what have you done for us!
A good question. What HAVE they done? Waylaid innovation, crushed entire companies, lied in public about it. Lied in court, too. Lobbied itself out of court, out of taxes, out of most everything. They can, they're sitting on a veritable hoard of blood money. ``Perfecting technologies''? Embrace and extend, they call it. They know about interop, alright. They're experts at subtly breaking it. Even their ``document format standard'', standardised through force-feeding it to several standards organisations like you stuff a goose (and breaking one while at it because their shills no longer show up for other meetings), doesn't have a single compliant implementation. What's that? Yes, their own implementation isn't compliant with their own standard. In the long string of such moments, this is a particularly crowning moment of micros~1.
So what was the point here? Reminiscing about a dark and stormy past, hoping that by adding a rosy light the crimes against technology fade a bit quicker? Are they trying to divert attention from their latest fsck-ups, like volume licensing? What shoe are they about to drop now?
"Even their ``document format standard'', standardised through force-feeding it to several standards organisations like you stuff a goose (and breaking one while at it because their shills no longer show up for other meetings), doesn't have a single compliant implementation."
You mean like Open Office isn't fully compliant with ODF 1.2? And I've always wondered why Elreg never covered the political machinations involved in getting the OSI to adopt ODF as a "paper standard", with no real world implementations, purely for the purpose of scuttling Microsoft.
...that because of flaws in OOo's ODF (initiated on OSI invitation, if memory serves), micros~1's retribution to that by astroturfing and shill proxying and everything else they've done is therefore excused? I think I'll trust bistromath before I trust your take on the bill.
Besides, micros~1 was never prevented from getting a copy of the standard and implementing it, or even from co-operating with OOo to work on ODF. It makes perfect sense for OSI to invite OOo to start work on ODF, them being an open source shop and thus used to open specifications, for the express purpose of coming up with an Open Document Format, without assuming malice. But please show how OSI was malicious towards micros~1 in doing what they do best, namely stanardising things and publishing the results. Your special math and all.
@serendipity: Mono is a nice idea and Microsoft seems to be helping it along.
But "open" does not apply. The patent pledges given by Microsoft are restricted and do not apply to everybody. For example, to the best of my knowledge, you can't do a commercial application with Mono unless it's running on Novell/Suse.
If MS wants to spur open .NET and Mono then do it. What it is doing now is the same old suck-em-in and extinguish.
@Steen Hive: As far as I am aware (and I've looked), the patent pledge from Microsoft covers non-commercial use, and commercial use on Novell. There might be more but these are the only ones I found.
If you're saying that Unity has one too, then great for Unity.
My point is that if a company develops and sells an application based on Mono, then both the company and its customers are at risk of a parent infringement lawsuit from Microsoft. If you believe this not to be true then provide the details
Did I fall asleep while El Reg lost its spine?!
"Microsoft's spent decades perfecting identity, directories, presence, and management that underpin things like location-based services in Exchange, Outlook, and Windows."
To excrete such a sentence with nary a drop of bile is a grotesque blot on The Register's once-proud escutcheon. That's a frickin' _vulture_ up there, innit?!
Save the puffery for elsewhere, please.
ignoring the fact that you open with the assertion that Amazon as based in the Valley (they're just down the road from Microsoft in Seattle) and end with referencing his 34 year history at Microsoft (and ramble on about location based services in Exchange!) this whole thing reads like some fluff peice put together to try and create a public persona to lend him some credibility.
And look how well the MySpace deal has gone as that spirals into the dust
Lewin has had zero public profile but since they exported Don Dodge to the Chocolate Factory I imagine they desperately need a new public figure to speak at events and host panels and glad hand the fellow fake tan crowd
Slow news day?
Oft-repeated wishful thinking from Microsoft fanboys: "Meanwhile, Apple has become re-invigorated, partly with the help of a cash infusion from Microsoft in the late 1990s."
Check any reliable history on the event. Microsoft was allowed to make a pittance $150M investment in Apple as a sign that Microsoft planned to continue support for Macintosh, so that Mac users might not be frightened away from Microsoft Office.
Apple didn't need the money. Microsoft made a huge profit on the deal both in stock and Office sales.
.NET sitting on top of OS (with security, performance, breeches, ... updates)
OS sitting on hardware (with security, performance, breeches, ... updates)
hardware (with security, performance, breeches, drivers ... updates)
Yes, all very voltage and power intensive yes?
Now how about .NET being hardware based (or at least the founding principles upon which it is based)?
"Well er let me see...oh that's it, how about Linux!! Check out the Mono project John!"
Neatly handled by the AC who posted after you I think.
"Oh and if that doesn't float your boat, you could always write ASP.Net web apps. They'll run in all the major browsers across all the main platforms."
Active Server Pages I presume. That would be where the server executes the script much in the same way the 1st web servers emulated the web page then copied it wholesale to dumb terminals (good implementation tricks never get old but they can get forgotten) . So *obviously* they should run on *any* browser. Which web servers support that?
Seriously I do get that some people like some of the features that .Net has given them but would prefer if they could host it on something else. I can certainly understand MS not wanting to do this themselves.
Where I and the Mono/Moonlight developers part company is that I can quite easily understand why MS would *never* want it to happen. Not that ehy won't do it, that they will take active steps (or should I say pages) to prevent it. However I had thought MS was *compelled" to publish details of all Windows interfaces, which along with the developer kits should list enough API entries, return codes, headers, data structures and overviews to make the port *fairly* easy.
Basically Linux is code, .Net is code. Different calling conventions, data structures etc. If you're looking at a different implementation language I could see big problems there, but as a clean room implementation you don't *have* the .Net source code to begin with. You're mapping (or even easier) duplicating data structures under one OS in another. then you're mapping the I/P and O/P specs from each API to your new. Lots of bulk but the first cut should be eyeballing the D/S and API calls in *detail*, then developing a set of replacement macros to break the back of the problem, before humans start chomping on it.
I do wish the developers well. They seem to want Moonlight/mono to be free, in the sense of freedom.
They (you?) might like to remember that freedom (especially where *any* large corporations business is concerned) is *never* given, it is taken.
A quote from the story: "Google has emerged as a centrifugal force for the web, society, and development." The sentence does not mean what the author thinks it does. Possibly he meant to write "centripetal force". Or, less likely, "centralizing". But certainly he is misusing "centrifugal".
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
