Fog Creek Software
Discussion Board




Another reply to "How MS Lost the API war"

http://www.longhornblogs.com/robert/archive/2004/06/18/3731.aspx

"...IMO, Joel just doesn't understand where Microsoft is going, which is as much his fault as Microsoft's. They've done an extremely poor job of communicating their long-term strategy..."

I haven't heard of this guy before, but it's a pretty decent little response.  He doesn't address the web app angle, but fends off some of Joel's salvos fairly effectively.

Pickle
Wednesday, June 23, 2004

So in essance can you summarize what the fuck he's say?

abusydude
Wednesday, June 23, 2004

This guy is a frickin intellectual titmouse compared to Joel. OH MY GOD.  He's a Microsoft MVP?  Well that settles the argument.

"IMO, Joel just doesn't understand where Microsoft is going, which is as much his fault as Microsoft's. They've done an extremely poor job of communicating their long-term strategy"

Yes we're all too pfuckin stupid to understand where Microsoft is going. Once Microsoft communicates and pounds it into our stupid heads, where exactly we're going, we'll follow along like stupid titmouses.

Microsoft's "LONG-TERM STRATEGY" is to release a new pfucking API every 26.7 days. Fire and motion, ya keep up or ya die. Well guess what the party's over. Joel's right, Microsoft lost, oh you say Longhorn will fix everything?  OH I CAN'T WAIT, LONGHORN IN 2007.

.
Wednesday, June 23, 2004

Fuck .NET. We are not risking our multi million dollar backend app simply because a 27 year old kid thinks it's suck a great idea to do all the well-crafted code in C# or Bill# or whatever the fuck is the new trend at MS.

Longhorn, my ass. We are happy unless, those mindless inexperienced assholes in MS can prove, logically not emotionally, why we should be doing this.

abusydude
Wednesday, June 23, 2004

Who is "Joel Splosky"? This guy even spells Joel wrong/

Matthew Lock
Wednesday, June 23, 2004

Actually, he spelled Spolsky wrong.

.
Wednesday, June 23, 2004

Interesting that anyone could think running Visicalc means that Win32 remains alive and well in Longhorn.  Visicalc hit hardware, all the Visicalc demonstration proves is that the virtualisation still works.

Nothing to do with APIs.

Simon Lucy
Wednesday, June 23, 2004

I think he is only 22.

Jackass
Wednesday, June 23, 2004

Way to nitpick and avoid what he said.

The Visicalc thing is there to make a point.  The point being that all of your Win16 and 32 code will still run. He also points out that all your WinForms code will still run.  XAML is in addition to all the previous methods of app development, not in place of it.

Oh, and if you bother to actually read the last paragraph, his pint is that MS has done of poor job of explaining what its plans are, not that we're too dumb to understand.

Now go back under your bridge.

Steve Barbour
Wednesday, June 23, 2004

pint ==> point

Although a pint or two would be nice.

Steve Barbour
Wednesday, June 23, 2004

Visicalc may be being used to make the point, but its the wrong application to use for it.  Visicalc used neither Win16 nor Win32 APIs, it only marginally used the DOS API most everything hit the hardware.

I don't particularly disagree that Win32 will exist as some kind of layer, that isn't the point that's being made.  Its that before there was a fairly restricted set of APIs, in future there will be many and they will compete with one another.  I made the point a few days ago that Microsoft is competing with itself now.

So, its no troll.

Simon Lucy
Wednesday, June 23, 2004

Nice reply . . . Robert is "dead on" w/ his retort. Well done!

Anon
Wednesday, June 23, 2004

Written well enough, and with a couple of good points, but ultimately I was left with the impression that the author is inexperienced in real software development (and it's a good thing that his input is limited to "found a seat beside someone with a name", though he implies that he's providing tactical direction for Microsoft): "Throw it all out! Let's start from scratch!". Uh..okay... The comment on the "unilanguage Java" was telling (there are a variety of languages that target the "Java" IL, though much like VB.NET and C# the design of the target often means that the differences are akin to find/replacing { with begin, } with end, etc - syntactical sugar).

Ultimately in this debate, like many, personal motives appear through the fog of rhetoric.

-"Oldtimers" that don't keep up who berate all this new fangled technology, and deride it as unnecessary and irrelevant.

-Younguns that never could figure out COM, and who have limited knowledge of COM/MFC/Win32, immediately hop onboard the revolution train. "Throw it all out!"  This is self serving because it puts them, they believe, at an equal footing with more experienced programmers, and allows them to be first-wave pioneers in something.

-Rich client developers that are addicted to their component sets and way of doing things, and for whom "DOM" and "DHTML" are frightening concepts, talk about the purported mass migration from web apps to rich clients (despite every industry metric indicating the opposite, but all you need is to repeat it to make it true).

-Thin client developers that have limited rich app development who proclaim that everything should be a web app...

And so on. Blah.

Dennis Forbes
Wednesday, June 23, 2004

I wanted to add my comments but Dennis nailed it down perfectly.


JD

JD
Wednesday, June 23, 2004

Simon, the bridge comment wasn't meant to refer to you, although in re-reading my comment it did look that way.  Sorry.

Anyway, I think Visicalc was great for demonstrating a point.  You are correct, it doesn't really show how well Avalon will supoort Win16 and 32 etc, but it was perfect to demonstrate the point that you will still be able to run all your old stuff on Avalon, even old stuff that no one in their right mind uses any more (I apologize in advance to all three people who are still using Visicalc for DOS). If Visicalc is going to run, then there's no reason to think that my Whizzy Win32 app won't. 

You are correct in that MS is now competing with themselves though.  That's what happens when your products really are "good enough".  There's no compelling reason to change.

Personally, I look forward to having a huge variety of tools to produce in, but I can see the downside of excessive fragmentation.

Steve Barbour
Wednesday, June 23, 2004

backward compatibility myths.
*Some* things always get lost on Microsoft OS upgrades.
For example with WinNT/(and Millenium?) they lost DOS applications that did raw device access.

Also, Visicalc is a real mode DOS application, it has nothing
to do with Win32.

Michael Moser
Wednesday, June 23, 2004

Joel:s "How MS Lost the API war", is an outcry that says, "My God not another framework!". And the reply is the usual "You don't understand it."...

MyNameIsSecret();
Wednesday, June 23, 2004

The ability to run Visicalc is impressive (if weird) but it only demontrates that you can run old apps. It does not demonstrate that you can recompile them.

When VB came out, many companies created specialized controls for it. After a while, many of these companies went away. The old application still ran but you could not recompile it.

Win32 will likely be provided for a very long time but will you be able to support applications that use it?

njkayaker
Wednesday, June 23, 2004

A major point in Joel's article is: the reign of thin-clients is arriving, so API's are becoming less relevant.  This has nothing to do with MS making mistakes, so I find it amusing all this arguing about whether MS is or isn't doing the right thing.  It is somewhat irrelevant what they do, says Joel, they *can't* retain their dominion in the world of applications.  A strong argument, I think.  I can imagine in 10 years you'll be able to use some shmancy web site to write a book instead of Word 2014.

It's actually old news.  Had that Robert even started college before ActiveX controls became totally irrelevant on the Internet?

Brad
Wednesday, June 23, 2004

Active X isn't irrelivant, it is now sinister.  I think Active X is probably IE worst security hole, and could be its downfall.  All of these trojans, CWS and the like, continue to get worse, and hijacks are on the rise.  Banner ad companies that are failing are now serving trojans with their banners, turning many popular non-professional sites against their users.

Ok, a bit apocalyptic, but my recent hijack experience has me worked up.

Keith Wright
Wednesday, June 23, 2004

The point that I don't see in any discussions is this:

Let's say the .NET framework becomes pervasive on the client. (Not that much of a stretch.  Heck - Mono may even deliver Unix clients.)

Then as a developer, what do you choose to write your complex UI in:

* Native Code (no thanks - too many deployment hassles.  ActiveX=dead)
* DHTML (ick! slow, complex if you try to elevate to be rich-like, splash around your source code, etc.)
* Avalon (wait till 2008 until there's an install base)
* WindowForms - Good language(s), deployable from a web server (sort of zero touch), good long term support  - let's be real - WindowForms apps will run for a long while).

The least evil seems to Windows Forms.

Just Me
Wednesday, June 23, 2004

"Native Code (no thanks - too many deployment hassles.  ActiveX=dead)"
This goes against the "Never Rewrite" rule. All the knowledge of development and deployment of Win32 apps is hardly a "non-issue".
Everytime people spit on 99,9% of what is out there, I have a good time.

Dewd
Wednesday, June 23, 2004

Yeah - If you're heavily invested and you're not changng much - why bother.  The problems with native code that propell me towards .NET are:

* Needing to have Administrator rights to install - .NET so-called "zero touch" gets you partly over that
* Wanting to make new investments "future proof"
* Developer productivity (as opposed to MFC/C++ or complex DHTML)
* A sense of wanting to deliver state-of-the-art solutions and anticpating that at some point it will catch on and become demanded/expected. (i.e. Betting against Microsoft=bad - yeah,yeah - this is point is non existent - if Joel is to be believed)

Just Me
Wednesday, June 23, 2004

I can agree with his perception of MS for the most part, although he makes it appear as if no support for the Win32 or the current GDI model. However, I think he claim the an AMM is a huge productivity gain is ridiculous. Especially, because he comes off as that AMM will give you these gains in lieu of OOD/OOP. Like they are even in the same category. Give me break!

thehappychickenking
Friday, June 25, 2004

TheHappyChickenKing - If I could understand your post, I might agree with you. 

That aside, it does seem to me, as well, that Joel is insinuating that there will be no support for Win32 application development in Longhorn.

CodeMonkey
Friday, June 25, 2004

*  Recent Topics

*  Fog Creek Home