I'm developer for a small company (less than 10 people).  An important part of my mandate is to maximize value per invested development dollar.  I look around me at the software development industry and it seems that _everyone_ is flogging Web 2.0 apps and interfaces to systems.  Perhaps it's the latent luddite in me, or perhaps it's just comon sense, but my experience has been that it is SO MUCH easier to develop rich user experiences in rich client platforms as compared to web/thin client environments.  I understand that web apps simplify application deployment and (to some extent) debugging, but let's face it, Web 2.0 is an _attempt_ to provide rich client experience over the web... and this requires considerable extra effort to overcome the inherently stateless nature of web protocols.

 So, in considering focusing on a web UI for an application, aside from the insecure pissing contest that can arise from staying current with the latest technology fad, what true value is there for the regular smaller company whose business is _not_ directly based in IT, but is dependend on it (ie. IT is needed to get the job done, but the focus is delivering a non-IT service).

 Take OnTime for example.  Clearly these folks have put a lot of effort into their web UI for the application, and for a web UI, it's pretty good.  But in comparison to the Windows client, it's definitely clunky and inferior.  I bet you that he Axosoft team has put a lot of effort to make the web UI as rich and flexible as possible, but let's face it... it's a LOT more work to achieve an equivalent level of control and impact as you can in a Windows client.

 So, what gives?  Is it like cell phones, where the nifty ability to have flexible and portable communication somehow makes us overlook and accept substandard (compared to land lines) communications, where there are echos, dropped calls, we can get billed astronomical rates by the _minute_, etc., etc., etc. 

 Yes it's really cool that we we can create desktop-like experiences in Web applications, partly because the concept is so unusual and often takes so much EXTRA effort develop.  But in a situation where one is trying to squeeze as much value as possible out of every development dollar, does it make sense to develop web-based apps for a small audience?

Whaddiyathink?

 - john.

John, this is an excellent question. I'm sure lots of people struggle with the answer to this very question every day. Both from the standpoint of development (should we create a web app or desktop app?) and from the standpoint of product adoption (should we choose a web-based product or desktop product?).

I think the answer is largely dependent on organizational structure as well as application needs. Consider the answers to these questions:

Organizational Structure

• Does the company or team have appropriate IT skills to manage IT needs in-house? (no or low IT, choose web, great IT skills, desktop apps become more feasible)
  
• Is the team distributed geographically? (for distributed teams, web apps is the way to go)
  

Application Needs

• Is the application in question extremely UI intensive? (the more complex the UI needs, the better suited desktop apps are)
  
• Is there a significant need for collaboration? (web apps are generally better suited for collaboration)
• Is the application updated frequently? (frequent updates require more installs for desktop apps, so web apps might be better suited for apps that are updated frequently)

So these are some things to consider when deciding on a web 2.0 or desktop application.

---

Hamid Shojaee
Read the Ship Software OnTime Blog
Axosoft, LLC

Hamid,

 As always, I appreciate your well thought out response.  Perhaps you're including client-server (remoted) apps in with the desktop category, but to be very clear, are you including this category in with desktop apps?  I mean with technologies like ClickOnce in MS-land, and I've experinced similar things with Java apps, the updating and distribution issue seems marginal.  I suppose one of the things I was curious to get feedback from others on was, my sense that that Web UI's take considerably more effort to develop a desktop-like experience, and even after all that effort the user experience is still not as rich and robust as Windows UI's.  I take OnTime as a great example.  Clearly you folks have put a lot of energy into the richness of the Web UI, and it's pretty damned good... for a Web UI.  Yet it is still not nearly as nice, smooth, solid, and convenient as the Windows client.  It allows people to get the job done, but I have found that it frustrates a number of users because its less predictable and harder to control than the windows client.  It's clear given your market that you absolutely need to develop a Web UI, yet it seems to me that it is significant that you also have a Windows client.  It's not really viable to only have the web client, and I think that's because of the usability issues I have mentioned.

I also question the premise that web apps make for lower in-house IT requirements.  Clearly, web apps are superior in that you don't have to worry about distribution issues, and desktop compatibility issues are greatly reduced.  Yet, the ease of use of an app's UI goes a long way to reducing support as well.  Especially when there are small numbers of users in a single location. 

Thanks for bearing with my deliberations. 

- john.

 

So a few more things to consider is that ease of use is not a feature of Windows or Web apps. It's a feature of design goals for either type of application. Clearly, if a windows and a web app could be equally good from the standpoint of user experience and performance, then the web app would be preferred, if for no other reason than to have the ability to use the application from anywhere using any computer. Technologies such as Click Once, java apps and Active Directory auto pushing of software to hundreds of clients only complicates matters more.

---

Hamid Shojaee
Read the Ship Software OnTime Blog
Axosoft, LLC

Yes, when the user base is sizable, web apps are clearly the answer to the challenges of distribution... no question.  But that issue aside, are you trying to tell me that the OnTime Web UI isn't quite as good as the Windows client simply because it's not designed well?  I find that hard to believe.

It makes more sense to me that it takes _much_ more effort to make a web client pretend that it can behave like a windows client, and even after all of that effort, the web client still comes up short, due to the inherent nature of the development context and the current state of development tools.