Fog Creek Software
Discussion Board




Why microsoft lost the API war.

Microsoft lost the API was because computers got faster.

Computers got really fast compared to the complexities of drawing a UI. So inefficient solutions like web apps are now possible.

Unless few fangled UI "paradigms" (ohh don't we all hate that word ) are developed which eat up tons of client resources then web apps will probably win the day.

Hence the new microsoft APIs, however I do not think that they address the core problem that a glass like UI, while cool, does not increase the user's performance.

Alain Hamel
Wednesday, June 16, 2004

Never one to dampen an interesting new point of view, can I just point out that despite Intel's ads that CPU's somehow affect browser performance, the actual bottleneck has been that wire going into the wall?

And right now only 50% of the US is on broadband (that is of the US that's online at all). Some analysts are theorizing that number may plateau at 75%. Less than half the hotels have broadband in guest rooms.

How many of the web application advocates live on dialup? (Honestly curious)

Philo

Philo
Thursday, June 17, 2004

Oh I was not talking about the round trip latency. That makes the entire system much worse of course.

I was simply observing that attempting to make a script driven UI would have been unbearably slow 8 or so years ago. However today such a UI system is quite snappy. Even if the UI code is horribly inefficient.

Alain Hamel
Thursday, June 17, 2004

One must also ask how many Windows systems that badly need updates are on dialup. 

Logical
Thursday, June 17, 2004

While I thought the overall article was interesting (and true), what I really learned today was that Ctrl+Enter prepends "www." and appends ".com" to anything in IEs address window.

Thanks, you just save me eight keystrokes every time I use IE!

ffoiii
Thursday, June 17, 2004

Come on Philo.  You know as well as the rest of us that the home user isn't what is driving the move to web clients.  It is corporations.  I'd like to know how much money is spent on software by consumers vs businesses every year.  I think that will show a different picture. 

The amount of money spent by people that have dialup in their cubical is probably pretty low as a percentage of corporate software sales, so considering corporate dial-up is probably not of strategic importance if you are as big as Microsoft.

Years ago I used to think that software was something I bought on a shelf, but what I failed to realize is that the software that does the real work that runs buisnesses has never been bought off the shelf.  It was sold to them by IBM or Computer Associates or DEC and now Microsoft.  To me there is a significant advantage for a company to put their expense reporting on a corporate web site vs deploying a thick client.  That's what is important. 

The only software that consumers pay for is what comes on their PCs and games.

christopher baus (www.baus.net)
Thursday, June 17, 2004

Sorry Christopher - I figured since Joel wrote the article and talked about gmail we were talking about home consumers.

Of course, if broadband is assumed, doesn't the complaint about the size of the .Net runtime go away?

Philo

Philo
Thursday, June 17, 2004

I never complained about the size of the .NET runtime. I don't even mind .NET.  There is a good chance that it will open the market as .NET is much easier to clone than Windows as a whole.  I just think there is a significant benefit from web deployment vs fat client deployment.

Java never worked that well on the client, I don't suspect .NET will either.  Until you see something in million lines of code range running 100% managed, I think it is unproven.

christopher baus (www.baus.net)
Thursday, June 17, 2004

If Visual Basic was such a best seller, where were the applications? I can't think of any major, successful applications that were built on VB. I know it sold a lot, but did it sell mostly to hobbyists and shareware writers? It always seemed to me that any shareware program I downloaded that required the VB runtime (which I usually already had so download time didn't influence my opinon) was generally bad. You know, get "get rich quick" sort of app that plagues some platforms (like palm). A depth of field calculator or something that the author wants you to spend $30 registering.

Seems like most commercial apps that really went places were built directly on win32; is that not the case? Mabye VB made big inroads in the custom programming space, but most programs I've run across there were built on top of Word, Excel, or more likely Access (I realize those are still VBA, but they require MS office don't they are not really standalone). Maybe I'm just missing some big apps that were in VB, or ignoring some market sector altogether?

Rob Meyer
Thursday, June 17, 2004

You're missing a market sector. A lot of small tailored client-server systems were made with VisualBasic + Access and things like that. A database and a quick UI to support soneone's work.

fyi
Thursday, June 17, 2004

Hello Joel, it me Joel...

Any way...I port Java's Object library to C++, 90%. Its doesn't use garbage collection, it uses reference counting. Its super fast compared to java and all the classes are the same. It works with VC++ 6.0 now. I don't think it would be hard to port; i'm planning on doing it anyway.

If anyone would like it, email me.
thanks

Joel Floyd
Thursday, June 17, 2004

> If Visual Basic was such a best seller, where were the
> applications? I can't think of any major, successful
> applications that were built on VB. I know it sold a lot, but
> did it sell mostly to hobbyists and shareware writers?

How about Joel's CityDesk?

Matthew Lock
Thursday, June 17, 2004

Hi Joel,
You said you like rich clients. What do you think about Macromedia's new product, Flex
http://www.macromedia.com/software/flex/

What could be the impact of such product ?

Ovidiu EFTIMIE
Thursday, June 17, 2004

> If Visual Basic was such a best seller, where were the applications?

Rob, Visual Basic was huge for in-house corporate and consulting development from about 1993 on. Every large corporation running Windows, which was just about all of them, would have significant VB development departments. Sometimes there would also be different groups that did C++, or something wierd like Smalltalk, but they would all have some VB.

Consultant
Thursday, June 17, 2004

Joel paints scary and very realistic picture.

I'm personaly hoping for the emerging winner will be "fat client on web channel" hybrid approach. People will start with what accessible (0% install reqs). But as soon as people will get comfortable with particular new technology, they will again go for ease of use, speed, perfomance and whole "realtime 3D fish aquarium when idle" thing.

google got inside our houses as web app, but for practical use its "fat client" toolbar. same thing with blogs. it usually starts out with habbit of  visting couple sites. It ends up in very fat and outlooky NewsGattor or somesuch. whats even more funny: blog site is rapidly becoming just a host for "RSS - subscribe!" button, less and less people actually are bothering to visit the sites:
http://avc.blogs.com/a_vc/2004/05/its_the_feed_st.html

Joel just nails it 100% on the head regarding MS keeping IE/DHTML in dark ages. However i think inovation just like water will flow around logjam. If MS won't let us make rich-client apps inside IE, we simply will start making fat-clients which consume proprietory HTML/XML streams instead of IE. Have a basic functionality web app site to get people interested, then give him fat client to really deliver on all your promises.

random thought: gmail as outlook killer? google goes gmail as planned, get X million users in for free. adds calendaring/scheduling/on backend and release "gmail toolbar" - which of course would be just a fat outlook-like client.

MaxS
Thursday, June 17, 2004

  See what Miguel de Icaza has to say in that connection. A bit different, IMHO.

doesn't_really_matter
Thursday, June 17, 2004

Let's not also forget Java Webstart, which allows you to run rich clients based on a JNLP file (and several other web-based resources which are all downloadable).  This is a very powerful technology that combines the accessibility of web apps with the flexibility of rich client UIs.

Adam Rice
Thursday, June 17, 2004

"
I'm personaly hoping for the emerging winner will be "fat client on web channel" hybrid approach"

Bandwidth will limit that.

"blog site is rapidly becoming just a host for "RSS - subscribe!" button, less and less people actually are bothering to visit the sites: "

And this is what Scoble and Microsoft hope.  Make the browser irrelevant.  RSS for all.

Quad
Thursday, June 17, 2004

CPUs get faster, then JavaScript runs faster.

Bandwidth get broader, more meat to the GUI.

Web standards among browseres is getting better, no more time wasted on trying to get the answer from 6/3 right in every browser ;)

Theses thing is what have been lacking in web apps the last 10 years

I hated JavaScript before, because of it was slow
today I am amazed about what it can do.

Thomas
Thursday, June 17, 2004

Quad writes: "Bandwidth will limit that."

Depends on how much data is going back and forth.

Take a very popular fat client app: iTunes with the Music Store.

Downloading songs, of course, is bandwidth intensive, as is downloading the 30-second sample clips. But the web-like UI and visuals probably would work reasonably well on a slow connection, as would the text data used in the hierarchical browser view of the store.

Likewise, Quicken works pretty well on a slow connection. You don't need much bandwidth to download your transactions or the latest quotes for your portfolio.

Jon Hendry
Thursday, June 17, 2004

*  Recent Topics

*  Fog Creek Home