Why lock ourselves in a silo?

22 November 2010

I recently got an email from a customer with whom I chat every now and then. Paraphrasing:

I just got my new Windows Phone 7 [he didn’t say which device it was] and I’m very impressed with how good it is. I’m surprised that DevExpress hasn’t been saying anything about creating controls for it. I think this is a huge market and a lot of Windows developers now have a mobile platform they are more comfortable developing for.

Given that we haven’t had our annual summit yet (it starts next week!), it’s perhaps not surprising that we don’t want to talk about our plans for WP7 yet. But, for some reason, I felt like playing the Devil’s Advocate that day and wrote a long reply, and it’s worth quoting here to engender some discussion:

The biggest issue is I don't see Windows Phone 7 being so big that we have to create native controls or libraries for it. Remember: the size of a market for applications for platform X is not necessarily indicative of the size of the market for controls for that platform. Yes, there’s some correlation but it’s not a slam dunk.

Also WP7 is only really appealing to developers like ourselves or our customers. It’s not particularly appealing to developers that use other phone OSes. So, yes, we do hear a lot about WP7: we're sitting at the business end of the Microsoft megaphone after all. (And remember Windows Mobile 6 was also “a mobile platform that Windows developers are more comfortable developing for": it used a Compact .NET runtime and it was only a success in a narrow niche.) But it's way early days yet to determine whether it's going to be big for retail customers. A far bigger retail market in my view is Android. Or iPhone. Or, shudder, Blackberry.

But the problem that I see is there's a plethora of phone OSes. Which one do you choose to develop for? Choose wrong and there's a problem. (And we go back to my CTO Message about hindsight and foresight.)

A far bigger market in my view is the one for web apps for phones. All modern (i.e., this year?) phones have a standards-compliant browser. They all have a damn good JavaScript interpreter (and they're getting better at it). Why not write web controls (ASP.NET MVC extensions? Mobile client-side UI controls?) that target them all?

You could say we at DevExpress pretty much write controls for business apps. Unless a business or a company can lock down which phones can access its systems (and here I'm thinking of my bank as an example), they will have to write apps for all of their customers' phones: iPhone, Android, Blackberry, and, OK, WP7, but in all honesty they only want to write it once. Hence they’ll write a mobile web app. That’s why I can currently go to WellsFargo.com on any smartphone I can get a hold of, and it looks the same: a nice mobile version of my online banking accounts.

(Of course, the same does not apply to games for phones. There the requirement is for great graphics, smooth animations of sprites and backgrounds and what have you, and native programming is the only way to go. But we don't even pretend to play in that space.)

And what’s wrong with MonoTouch all of a sudden? Talk about “a mobile platform that Windows developers are more comfortable developing for". (Well, OK, apart from having to use a Mac. Which I do these days.)

So, what do you think? Am I completely off-base, dreaming in la-la land? Or have I gauged it correctly: for business apps, the OS doesn’t/shouldn’t particularly matter?

9 comment(s)
Michael Thuma

From a technical perspective devexpress will know it better if there are show stoppers. MS is simple, as they have no strategy decide for the way they implement their "Client"software for their business apps with. Look what it is missing ... when we come to Silverlight.

I see this independent from the web driven approach. If it is possible to replicate data for business apps and combine it with HTML controls this would be my preference and this is think is what HTML5 + the browsers will have to solve.

22 November, 2010

As far as my experience can tell, web mobile LOB apps sucks. Their only advantage over native apps is intercompatibility between mobile platforms, and it's not even perfect yet (though I have to admit it's getting better since the iPhone came out).

And even if JS wasn't an issue, it's still hard to provide the end user a rich, highly responsive web application. It uses more bandwidth (some are still paying quite high fees for mobile web surfing) and, likely, more precious battery life.

With more and more apps being added to dedicated platform's "app store" equivalent every single day - thus meaning more and more programmers having mobile development experience -, I have serious doubts on the future of web-based mobile applications.

Anyway, why is it so scary to consider maintaining two, three or even four plaform-specific LOB applications UIs, given that those are much less complex than desktop UIs and that the business logic should logically reside on a server exposing standard web services?

22 November, 2010
Peter Thorpe

I don't disagree on your market size estimations and it's difficult to say if Wp7 will be a success or not although I love my Omnia 7. So I can't say WP7 controls would be a commercial success for you for or not.

But then I don't think the web is the answer. The reason iPhone apps are so popular and not just for gaming is people aren't happy with the poor user experience both in speed and incompatablity as Crono said. Apps are so much more responsive and usable, which is what people want. When I had an iPhone I used ebay, IMDB, wikipedia, flixter (rotten tomatoes), google maps,  you tube... apps to name a few not in the game space. Not native or phone optimized websites.

Using MVC controls for mobile development would in no way drive me if I were a new customer to any mobile development using DevExpress controls. With the limited screen space and limits of the web I would just use freely available controls to create a clean UI if I wanted it as an optimized site.

22 November, 2010
Christopher Todd

I'm no fan of Apple by any stretch of the imagination; however, I had to return my WP7 and get an iPhone. I tried hard to like it, but I failed to convince myself that I needed to be strong and suffer it for a while. I hope the best for them, but until they realize that style only goes so far and simple function is everything, they will fall behind Apple. It always seems like it takes MS 4 versions to get it right. Why does WP7 not have a virutal volume control? You would think that would be the first thing they built! Nope! And to make you suffer even more, they make everything function off of the same volume setting. Yikes!

Personally, I think you'll already have way too many products to support as it is. Until MS can show everyone they can swim the full length of the pool and back, I wouldn't waste any time with them.

23 November, 2010
Michael Thuma


>> should logically reside on a server exposing standard web services?

But not in a cloud app. The logic should reside there but the standard web service no chance.

23 November, 2010
Christoph Brändle

there will be rich clients and wcf/websockets, and I doubt WP7 will not be in the game. the of course more elegant solution would be rich webapps on every phone, this dream wont die, but it wont be reality. no need for WP7 controls now, but tomorrow.

23 November, 2010

I have an iPhone and I use many apps and almost no web sites.  Websites suck.  The Chase banking app is wonderful to use, the Chase mobile website is excruciating.  The facebook app works nicely, going to the website is near useless.  I also use a blackberry and anything web based is terrible while the apps work well enough that I use my phone to post pictures and to read responses.  I wouldn't bother with the web....

I understand that you have a tough call to make in picking a platform so you only have to do things once, picking the web on phones feels like a losing proposition to me even if it is the least work.

If you could do something that enabled Win7 apps and iPhone apps to share a common code base (e.g. support MonoTouch) it seems like a win, or at least a tool that could make compelling apps and provide a crossover for developers used to a .NET programming model.

23 November, 2010
Bassam Abdelaal

Like Crono , I also believe web mobile LOB is very poor , but that’s not the issue , that’s argumentative subject , the real issue is that MS did not choose it as the application platform for WP7 , it used Silverlight instead , same for Apple they did not choose web LOB either and their store have more than 250K applications , WP7 is having 2700 applications in one month so there is clearly a momentum building here for it and apparently MS will continue enhancing it very strongly, other component vendors believed in the platform and started their beta control set for WP7 already, I think its not appropriate to mention their name here, so if the platform flies – and many think it will – then DevExpress will be behind those others believed in it and took the risk, but at any case web mobile LOB is surely not the answer.

24 November, 2010
Daniel Dacey

Like the other poster, I don't believe the web is the answer for mobile solutions.  I have an iPhone for my personal use and run native apps in preference to starting up Safari everytime.

Now I also have an Android and WP7 phone in my home and those family members also prefer native apps too.

I also developed and supported a web based iPhone app last year and hated every moment of it. No I don't want to go there ever again if I can avoid it. The web is a terrible programming environment, compared to native app development.

As for the vendors I like the WP7 phone, but I have to say among'st work colleagues it is certainly not a hit. Ironically the secret to making WP7 successful is to make sure every Android phone runs Silverlight. Then you can easily support a much larger market (Android) while still supporting and incubating a much smaller market (WP7).

Of course this would mean MS getting out of it's own way, which after recent Siliverlight announcements, looks highly unlikely.  They still don't realise how out of touch with the .NET development community they are. How else can you explain the increasing complexity of .NET with no real benefit to many corporate clients or developers and the rudderless direction of Silverlight.



24 November, 2010

Please login or register to post comments.