.Net: 3 Years of the 'Vision' Thing

[asj]

interesting read from eweek about the dying hope surrounding .net

http://www.eweek.com/article2/0,,1184728,00.asp

The end of last month marked the third anniversary of Microsoft's launch
of its .Net strategy, which executives such as Chairman and Chief
Software Architect Bill Gates said at the time was a "bet-the-company
thing." But three years later, reactions are mixed as to whether that
strategy, along with the vision that accompanied it, has played out as
the Redmond, Wash., software developer had hoped.

Rob Helms, research director for Directions on Microsoft, a research
company that tracks Microsoft, in Kirkland, Wash., said the .Net
initiative described a vision for how software and the Internet would
evolve; a new platform for software development that supported the
vision; and a new business—application hosting—that would drive future
growth for the company.

"Three years later, most of the hopes behind the .Net initiative have
not been realized," Helms said, adding that .Net has now almost vanished
from Microsoft's vocabulary.

On the thorny issue of the .Net platform versus the Java platform, Helms
said the slow growth of Web services has prevented Microsoft from
driving adoption of the .Net platform or giving it a leg up on Java. The
..Net platform itself has been hampered by immature Web service
standards.

 

[asj]

i meant "dying hYpe", not hope, but i guess that works just as well ;-)

 

[Roedy Green]

"Three years later, most of the hopes behind the .Net initiative have
not been realized," Helms said, adding that .Net has now almost vanished
from Microsoft's vocabulary.

The problem is, the industry can't bet the farm on a technology that
Microsoft might dump. With Java, at least if Sun loses interest,
there are plenty of other parties who could take the ball.
Java has been fiercely multiplatform and multivendor since day 1. MS
has been fiercely proprietary.

Gates underestimated the impact of the Internet back in the 90s. I
think he has done it again. He misunderstood the need for software to
run on a wide variety of platforms from tiny handhelds to humongous.

I don't want to see .net die. The .net folk seem much more willing to
cater to the needs of application programmers to make their life
easier, even if it complicates the lives of the compiler writers. .net
is a good spur to Java's continued evolution to be more application
programmer-friendly.

 

[Eric Gunnerson]

The problem is, the industry can't bet the farm on a technology that
Microsoft might dump. With Java, at least if Sun loses interest,
there are plenty of other parties who could take the ball.
Java has been fiercely multiplatform and multivendor since day 1. MS
has been fiercely proprietary.

Gates underestimated the impact of the Internet back in the 90s. I
think he has done it again. He misunderstood the need for software to
run on a wide variety of platforms from tiny handhelds to humongous.

I don't want to see .net die. The .net folk seem much more willing to
cater to the needs of application programmers to make their life
easier, even if it complicates the lives of the compiler writers. .net
is a good spur to Java's continued evolution to be more application
programmer-friendly.

The article is referring to the "Let's brand everything with .NET" marketing
frenzy of a couple years ago. That, thankfully, has run its course.

It's not referring to the .NET Platform and Frameworks, which are still very
much alive and talked about throughout the company. Longhorn will bring a
new set of managed interfaces to the O/S.

 

[asj]

Were was Java after 3 years? NOWHERE.

actually, by 1998, java was pretty much heralded as the next coming of
the messiah....it was being slotted into almost everything possible,
from the ill-fated NCs, to smartcards, to set-tops, etc, not to mention
the fact every second website had a really annoying, large applet
choking people's 28k modems. it was also trying to make it to the
enterprise at the time, with much derision and skepticism from some
quarters...and, oh, yeah, several idiots like corel were porting entire
office suites to it, even though the api's and jvms to handle such large
projects were not yet mature enough.

 

[Chris]

i agree with you here, that microsoft has done a MUCH better job to make

life easier for developers. and it HAS forced java to make some changes,
even including many initiatives to make java almost VB in terms of ease
of programming (how long did it take the java vendors to think THIS up?)

Although on principal I would agree with the sentiment, there could be
a problem here: if developing Java apps becomes easier, any monkey
could hack something out and then we'd all have to reduce our rates to
the level of VB developers - not good!

Perhaps we should start petitioning Sun and Java-based vendors to
reduce functionality & user-friendliness?



- sarge

 

[Jay Glynn]

actually, by 1998, java was pretty much heralded as the next coming of
the messiah....it was being slotted into almost everything possible,
from the ill-fated NCs, to smartcards, to set-tops, etc, not to mention
the fact every second website had a really annoying, large applet
choking people's 28k modems. it was also trying to make it to the
enterprise at the time, with much derision and skepticism from some
quarters...and, oh, yeah, several idiots like corel were porting entire
office suites to it, even though the api's and jvms to handle such large
projects were not yet mature enough.

Sounds like you just described .NET.

 

[asj]

Although on principal I would agree with the sentiment, there could be
a problem here: if developing Java apps becomes easier, any monkey
could hack something out and then we'd all have to reduce our rates to
the level of VB developers - not good!

Perhaps we should start petitioning Sun and Java-based vendors to
reduce functionality & user-friendliness?

you're too late...sun wants to grow the number of developers to 10
MILLION in 3 years using something called "RAVE", which would make
creating advanced functionalities in java a drag and drop thing....if
that happens the ratio of supply of developers to demand will be larger,
perhaps driving enterprise java salaries lower overall....

http://www.theregister.co.uk/content/53/31667.html

"Sun Microsystems has finally unveiled its long-promised development
environment to simplify programming in Java. Called "project Rave", it
is designed to bring drag-and-drop capabilities to programmers building
Java-based web services and applications."

"On a strategic and technical basis, Rave holds promise as a developers'
tool. Due in 2004, Rave is expected to compile Java code at the touch of
a button, and uses Java Server Faces to develop interfaces using
JavaServer Pages (JSPs). It is designed to import UI designs from
software such as Adobe Systems' Illustrator."

 

[asj]

Sounds like you just described .NET.

not that i know of...perhaps 3 years ago (when even gates described .net
as the be all and end all for microsoft), but not today, when it's
pretty much just another development environment for microsoft shops
that is cannibalizing older microsoft technologies like VB and COM.

 

[Mike Schilling]

It's not referring to the .NET Platform and Frameworks, which are still very
much alive and talked about throughout the company. Longhorn will bring a
new set of managed interfaces to the O/S.

Including one to flush from a FileStream to the disk? It's annoying to me
that I need unsafe code to do this today.

 

[John]

not that i know of...perhaps 3 years ago (when even gates described .net
as the be all and end all for microsoft), but not today, when it's
pretty much just another development environment for microsoft shops
that is cannibalizing older microsoft technologies like VB and COM.

Obviously you don't know what your talking about. Nothing unusual
there.......

 

[asj]

Obviously you don't know what your talking about. Nothing unusual
there.......

really? so where are most of the c#/.net people coming from? what with
all the questions from vb and other microsoft developers around here,
sure sounds like a mass migration of the herd from vb and other older
microsoft tech to c# and other .net.....and that sounds like pure
cannibalism to me.

unfortunately, .net has not dented the move of disgruntled vb users to
java either...what was that stat: 31% of disgusted vb users still moving
to java?

 

[Eric Gunnerson]

Including one to flush from a FileStream to the disk? It's annoying to me
that I need unsafe code to do this today.

Really? Does FileStream.Flush() not do what you want?

 

[Mike Schilling]

bring

Really? Does FileStream.Flush() not do what you want?

No, it does not. FileStream.Flush() flushes in-process buffers to system
buffers, but makes no guarantee that the system buffers are then flushed to
the disk. This risks information being lost in case of a system crash. The
only way I know of to force the physical flush is unsafe code:

FlushFileBuffers(fileStream.Handle);
...
[DllImport("KERNEL32.dll", SetLastError=true)]
static extern bool FlushFileBuffers (IntPtr handle);

 

[Grant Wagner]

bring

Really? Does FileStream.Flush() not do what you want?

No, it does not. FileStream.Flush() flushes in-process buffers to system
buffers, but makes no guarantee that the system buffers are then flushed to
the disk. This risks information being lost in case of a system crash. The
only way I know of to force the physical flush is unsafe code:

FlushFileBuffers(fileStream.Handle);
...
[DllImport("KERNEL32.dll", SetLastError=true)]
static extern bool FlushFileBuffers (IntPtr handle);

This seems like a non-issue. The system could crash between the lines which
flush the in-process buffers to system buffers and your code to flush the system
buffers to disk, and you'd still end up with corrupt data, how do you protect
against that? The disk controller could freeze or cause data corruption or loss,
how do you protect against that? The disk itself could experience a hardware
failure which would result in incorrect or corrupt data, how do you protect
against that? Worrying about this one aspect of data loss without addressing
every other possible event which could result in corrupt or incorrect data seems
pointless.

Flushing to the system buffer is sufficient, let the operating system take care
of moving the data to disk for you when it's convenient for it to do so.

If you're worried about power failure during the period between flushing to the
system buffer and that data being written to disk, get a UPS that will provide
sufficient capacity to last long enough to write any current system buffer to
disk (which would be a few minutes at most, more likely a few seconds).

 

[Pete Wright]

Well said Eric And good to see you popping up in the newsgroups from time
to time


--
Peter Wright
Author of ADO.NET Novice To Pro, from Apress Inc.


_____________________________

 

[Mike Schilling]

It's not referring to the .NET Platform and Frameworks, which are still
very
much alive and talked about throughout the company. Longhorn will bring
a
new set of managed interfaces to the O/S.

Including one to flush from a FileStream to the disk? It's annoying

to
me

that I need unsafe code to do this today.

Really? Does FileStream.Flush() not do what you want?

No, it does not. FileStream.Flush() flushes in-process buffers to system
buffers, but makes no guarantee that the system buffers are then flushed to
the disk. This risks information being lost in case of a system crash. The
only way I know of to force the physical flush is unsafe code:

FlushFileBuffers(fileStream.Handle);
...
[DllImport("KERNEL32.dll", SetLastError=true)]
static extern bool FlushFileBuffers (IntPtr handle);

This seems like a non-issue. The system could crash between the lines which
flush the in-process buffers to system buffers and your code to flush the system
buffers to disk, and you'd still end up with corrupt data, how do you protect
against that? The disk controller could freeze or cause data corruption or loss,
how do you protect against that? The disk itself could experience a hardware
failure which would result in incorrect or corrupt data, how do you protect
against that? Worrying about this one aspect of data loss without addressing
every other possible event which could result in corrupt or incorrect data seems
pointless.

You're right that being able to flush to disk isn't sufficient for building
a reliable transactional system, but it is necessary. And if you've ever
used a database, you've relied on that sort of flush working correctly
whether you knew it or not.

 

[Heinz Getzler]

i meant "dying hYpe", not hope, but i guess that works just as well ;-)

Maybe Microsoft feels that .NET is a failure because hailstorm did not
take hold and extend & embrace the entire web........

 

[asj]

Maybe Microsoft feels that .NET is a failure because hailstorm did not
take hold and extend & embrace the entire web........

microsoft can go kiss my fat as* for repeatedly trying to monopolize
everything. remember kiddies, the only thing worse than abusive
monopolies are people who don't care or do anything about it, because
then we all, as consumers, pay the price.

 

[Ben]

Here here!

Obviously you don't know what your talking about. Nothing unusual
there.......