Sunday, May 13, 2007

Blasts from the future.

This is a great spot for dotNet news:
If you're going to follow VCSY you need to know a bit of the .Net history in news =

If you are an IT manager trying to get a grip on Web Services and what you need to stay in the game, you DEFINITELY need to read the following from July 17, 2003:

begin article,14179,2914294,00.html
.NET: A guide for managers

By Jason P. Charvat
July 17, 2003

What is .NET?
.NET is essentially a set of software technologies designed to connect your world of information, people, systems, and devices. Therefore it’s an environment for building, deploying, and running Windows applications and services. A fundamental idea behind .NET development is the idea of common objects--objects that are accessible anywhere, anytime, and from any device. Also, .NET technology itself is based on XML Web services, which use standard protocols such as SOAP and XML data to connect applications and Web services.
end excerpt
end article

If you are a developer, I want you to think through what you're going to do in a world where your work is arbitrated into commodityship at an astounding rate. Once the managers catch on to virtualization and arbitration powers, you're going to be relegated to programmer farms to support the individual objects Microsoft will need to keep their islands of automation universe perking along... while the rest of the world does the next level up in abstracted application building... and kicking your ass with your own products.

begin article
Compact Framework Blues
Jonathan Goodyear, MCSD, MCAD, self-proclaimed Internet Bad-Boy

XPath is another victim of Microsoft's attempt to reduce the size of the CF re-distributable. In this online whitepaper, Microsoft's suggested alternative is "a combination of recursive and iterative searches against the Document Object Model". Yeah...I'll bet that'll perform well; Especially on a PocketPC device with limited memory and processing power.
end excerpt
end article

Yes, oh skeptical one, the article only references .Net in a mobile framework such as CE. but the very idea you would excuse .Net on that point outlines the problem with .Net. A different product has to be done and cobbled for different situations entirely... rather than delivering on the "any object" "any where" "any time" claim central to .Net marketing and central to any truly arbitrary framework. .Net in this incarnation was far from an arbitration framework much less a system using arbitrary objects. And the complaint that I would be unfair to .Net as a whole based on a selective view as "But this is .Net for a mobile platform" is either an intellectually dishonest feint or a "design only what they ask for" approach to engineering.

Arbitration removes uncertainty and provides machine readable and actionable governance in order to be autonomous. The truly arbitrated "object" may be applied by machine driven frameworks rather than a programmer having to decide which development kit to buy.

Now, look at the Microsoft concept and then take a look at the IBM concept:

Microsoft will not get virtualization capabilities out to public developers until 2008.

IBM has been employing virtualized components, applications, systems, frameworks... the works since 2001 in development and client businesses.

IBM will be providing automated service systems when Microsoft is first rolling out systems to be used by humans first. THOSE developers will be the ones expected to build out autonomous services once Microsoft's equipment comes online for them.

Why? Why should they? They will be competing with a race of technological giants that will flick them like fleas. Oh, Microsoft will still be cool selling, as it were, to their userbase (designers and developers) while their base fades into irrelevancy... enabling MSFT to buy them at a five and dime price and then join the 'autonomous services' paradigm around 2010... right about when they said such would be blooming.

Those Microsoft management types they are some smart, no?

Sure, there will always be programmers. I know the programmer has to get an object out there in the first place, but we are today virtually swimming in objects of every kind by any kind of proprietary re-invent of the wheel... and the best out of the pile will make it's way to the top of the application exchanges 'best of breed' for inclusion into new, future and legacy developments of the composite application.

There probably is nothing of substance in actual data processing done that has not been embodied in some perfectly usable body of cobol or fortran forty years ago. It's just you can't use that code unless you first virtualize it so it can communicate and interoperate with other code bodies ( look at US 7,076,521 and you can see how virtualization works with a client agent handling the mediation between 'you' and anybody else you have to hook up with and have intercourse with (don't be so shocked. data architects have been talking like this for years. it's all you people who demigraded the codewords.)). Once all the code bodies are integrated the overarching framework US 6,826,744 that glues and holds and implements this aggregation becomes the application. It's like cells virtualize in the goo and the body comes together as each part plays its part.

Do you SEE how far Microsoft actually is from this realization (expression + experience = realization)? Yes, they will get there eventually and it will be such a surprise to the rest of the world who forgot about what they were doing long ago so their public base will be impressed... eventually. That is if all these developers stay with .Net... or if .Net survives. A whole lot can happen... and each different track may be projected and somewhat predicted as we're in a very narrow technological pass from one paradigm to another.

Anyway, as with all other technological disruptions, "Thanks oh pilgrim prospectors and developers and programmers. Now, get off the settler's land."

We are leaving behind programmers in the VCSY paradigm whereas Microsoft in their past, present and future .Net paradigm has a vested interest in keeping the programmers plugged into the action and maintains that regime today with the divesting of functional integration with content/format (look and feel) GUI in one body of developers (designers) using expression and the marketed invisible line separating them from the designers (developers) in .Net Framework.

TODAY... NOW... Microsoft is making squeaky toddles over to a total user-defined experience, but they've rather chosen to mince steps with words and offer tools for Expression (content and format [look and feel]) and then end up having to issue the same tools (different names, of course) for functional assembly to the designers. THAT must have been a hard decision because I believe it put them in a do or die situation... the cutting of three key elements of their virtualization program essentially eliminates Microsoft being able to act as anything more than a Microsoft managed assembly of server units AND places hardware constraints on their versioning so they will always be a slave and never a master in an autonomous server bank. They will always need humans for specific gerrymandered tasks and thus will mask their true cost of implementation in their 'low cost software packages'. You will need many more developers per project than the 'high cost' application services resulting in many more problems, issues and cost than the web service as an application deliverables.

AND, when a composite application is built in the real world, there will need to be an army of developers to make that same thing happen in the .Net Framework implementations due to this set of limitations to virtualization with outside resources in software/hardware.

NOW we see why Software AND a Service are so important to Microsoft. They'll lose their developer base if they don't give them something to do and an artificial economy is better than no economy at all.

I don't think Microsoft intends to compete with virtualization. The question is why? To save their developer base? Are you kidding? If Mom and Dad and Junior could make Microsoft applications don't you think MSFT would slap that out there and let the developers swim on their 'let a professional' floaties?

I don't think Microsoft will be given a license to legalize.


Remember the dBase interface?

That meant you were supposed to tell the database what to do. Here we are twenty years later and we STILL have to tell the stupid things what to do. Now that the machiens will understand what a command line can convey, who needs the fancy shmancy graphics tools for objects? Why? Tell your computer what you want and get off the button pushing treadmill the graphic user interface demands. A whole lot of money for very little facility in this newly "virtualized to the gnat's pucker" world. Thanks for the memories, Microsoft. Pardon our dust.

EVERY element of an arbitrary distributed system must be ready to take into consideration the minimal platform... therefore, EVERY definition of a distributed architecture must be viewed as a "mobile" platform even if every last computer has a lead weight in the bottom for ballast so it can't be moved. the very nature of the uncertainties of network routing and network outage demand the marshaling of metadata and "location" is only one more piece of information.

.Net as it was originally built would not do those things. Why? Because .Net relied on SOAP and SOAP relies on all the intelligence and business logic reside in servers rather than in client agents at the site of the user activity. If .Net has gone 'the other way' and that is by client agents... well, buckarooos, not only is .Net infringing, but they have built many many products and client products that do thuswise also.

And, if Microsoft has been playing hardball the way they appear to ME to have been playing hardball (it doesn't take a conspiracy to make things happen) I would say in Mister Wade's eyes it's going to be the people who've worked loyal-like with VCSY these years who are going to be rewarded with a license and Microsoft and their minions and pinions will get no grease but fire.

Most software design is done with an eye toward cleanness under the reigning situation. "What will not take place" does not have to be designed for in most applications. BUT a general purpose application or a framework that's touted to be "all inclusive" can't take that track no matter how attractive the monetary rewards will be as you are stunting the platform for money... or lack of innovation or property.

You can bluff your way through when you're the one doing the delivery of services and applications as Microsoft did during this period. But, if there comes a time when the machines start distributing your work, you had best hope nobody slipped a small device on their desktop linkup or it's the proverbial monkey wrench in the gearbox.

As we move closer to an autonomous server and service mentality, Microsoft will appear more and more antiquated. Nothing personal, just an observation. It's already been happening in a small scale as a result of developers getting hold of Vista and asked "Is THIS what we've been waiting for?".

What happened Friday is about to enlarge that scale immensely.


1 comment:

Anonymous said...

BoB Jr.