Farewall to PaaS Provider, Coghead

Feb 19, 2009 by Sinclair Schuller

Coghead announced it was shutting it doors yesterday evening. For those unfamiliar with Coghead, it was a much talked about PaaS offering that offered an online application editor for rapid application development. They fell under the category of non-standard stack, a “4th generation language and runtime” (4GL) if you will. It seems that slow adoption coupled with the economic situation got the best of them. This is clearly a big problem for Coghead customers, and is indicative of why I don’t particularly like the “PaaS requires a new language/runtime/stack (pick and choose) because older runtimes and languages simply don’t work” statements. First, the notion that something like C# (or any .NET runtime language), Java, or Python can’t work is simply a problem with fundamentally understanding runtimes and what they can and cannot do. Second, new languages and stacks are generally dangerous unless they have MAJOR support from a large platform vendor. Why? Risk. Coghead is letting folks know they can access their data until April 30th, 2009. Unfortunately, data is not the application and any VARs, ISVs, or IT departments are up the creek without a paddle when it comes to the application itself. If the apps were built on some industry standard runtime, and the PaaS injected native SaaS into the applications, they would still retain the ability to run and use the code elsewhere (this is what SaaSGrid does). Last, some 4GL has a tough time leveraging existing code assets that your company may have invested thousands or millions into.

As for Coghead, it’s always a shame to see a startup go, despite our market approach differences. I wish the Coghead team the best of luck. Fortunately for customers, it didn’t take long for many of their competitors in the “WYSIWIG” DIY application platform space to come to the rescue:

1. Intuit QuickBase>
2. TrackVia
3. Caspio
4. TeamDesk

There have also been a few development firms offering to ease the pain as well:

1. Scio Consulting (Full disclosure, Scio is a partner of ours at Apprenda)
2. DeliveredInnovations

While it’s great to see that there are options for Coghead’s customers, it also causes me wonder once again why anyone would want to lock themselves into a full non-standard stack offering?  Granted, I have a bias as one of the creators of SaaSGrid, because our model is the exact opposite. However, our model is the exact opposite because we didn’t even see that as a reasonable approach to doing business and this is a clear example of why. A startup should focus on ensuring that it’s customers are safe and are not sinking time and money into something that cannot function outside of that runtime. If you’re building ontop of a runtime that does not have a natural “fail safe”, take a lesson from what’s going on here.

Is my fear of non standard languages overblown? Is reducing risk a better ‘key feature’ of a 3GL based PaaS, or is the fact that it can leverage existing software assets the more appealing attribute? If you are a Coghead customer, have you found alternatives yet?

The SaaSBlogs group on LinkedIn now has 1450+ members and it’s growing every day; make sure you are not missing out and join today.

read more

What PaaS Isn’t: An Application with an API

Feb 13, 2009 by Sinclair Schuller

For some time, I thought I was alone in battling the cloud taxonomy war. With SaaS, PaaS, IaaS, RIA, etc. being tossed around on a regular basis, I often find myself looking for clarification when it becomes obvious that someone’s definition and understanding of one of these acronyms is different than mine and those that share a similar view as my own.

Given my relationship with SaaSGrid, the one that hits closest to home is ‘platform’ and PaaS. I bumped into a great read by JP Morgenthal that attempts to build some understanding around PaaS. Morgenthal asks whether or not simply having an API makes a web application a ‘platform.’ Does it? No. As Morgenthal points out, it seems difficult to justify Facebook as a ‘platform’ if it doesn’t act as host to application code, but simply exposes an integrations API. Unfortunately, we see folks throw platform and PaaS around all too loosely. If we think about desktop applications, it would be awkward to refer to Microsoft Word as a “word processing platform” or an “office tools platform” simply because it has an API. Some do it on occassion, but everyone is comfortable calling Word an application. Furthermore, Microsoft never made silly claims like “Write the next Word on Word” or “All desktop software will be written on the Word API”. It has it’s place in the world. However, everyone is comfortable calling .NET, or Java, or the JBoss Application Container ‘platforms’ because beyond having APIs, they act as hosts to guest software. I would say this is critical minimum criteria in claiming to be any sort of platform.

My guess is that marketing departments found that using ‘platform’ makes everything sound bigger and better. It’s not a sales management application, it’s a ‘sales management platform’. It all reminds me of a bet by the Royal Chemistry Society: they are battling the marketing folks that have used the word ‘chemical’ as negative spin in hopes of boosting the image of ‘natural, chemical free foods.’ They basically claim that the word ‘chemical’ has been misused in a very harmful way, teaching people that ‘chemicals’ are bad. In fact, all foods have ‘chemicals’ and being ‘chemical free’ is an absurd claim created by marketing departments. ‘Platform’ is a little more ambiguous than this, but the spirit is the same.

Do you agree that platform and PaaS are often misused? Does it seem to you that the word ‘application’ is falling out of favor as a proper description for standard, non-host software?

The SaaSBlogs group on LinkedIn now has 1400+ members and it’s growing every day; make sure you are not missing out and join today.

read more