Here Kitty, Kitty… KittyHawk
Sun, Aug 1 2010 14:39
I was having a quiet coffee this Sunday morning, catching up on some emails, when some strange feline contexts took over my email. The first was a “MEEEEOOOW. LOL” post someone had posted on my facebook page. The next was something about Kitty talk. Was this the revenge of the LOL Cats ??
Whilst the “MEEEEOOOW. LOL” post was about a prank pulled on me last night, it turns out the “Kitty talk” was about a recent post by Mary Jo Foley speculating on a new RAD tool code named KittyHawk. Google shows some feedback to this suggestion: some for, some against (some almost violently so). The criticisms I read all seemed to focus on Microsoft Access or similar type applications. I think they’re wrong to liken it to Access as it wouldn’t make sense for Microsoft to release yet another competitor to Access (yes folks Access still ships today;) ). One thing is for sure there is a huge application development void that other companies, such as google, are looking to fill. It would make sense for Microsoft to also address this area.
If we look at where we are today, despite computers being faster, cheaper and closer and closer to being ubiquitous, software development seems to have gotten further removed from the end user. Even for professional developers, often the existing tools are still too slow in producing standard forms over data applications. Writing code to do transformations and validation has only indirectly progressed in the last decade, with little targeted directly at these end to end scenarios.
Some areas that seem obvious for improvement include:
+ drag and drop data binding: This has been close with WinForms, and now with VS2010, XAML has drag and drop binding. Still it’s not quite there yet; still demanding a greater knowledge than the task really requires.
+ data entry validation. Still today people argue what tier this should be in, and still there is no clear abstraction that lets the validation rules themselves be applied to different layers.
+ And from the UI, there’s still a lot of glue code needed for simple operations such as date2 has to be later than date1, even though that’s such a common scenario.
+ To have input that allows entry of new names yet shows completion lists with existing names
+ to have simple things like spell check
+ to have automatic separation of tiers, allowing ease of substitution of things such as data layers for in the cloud and back.
+ to have flexible drag and drop and type a little type of designers for data transformation without needing a degree in any particular data framework.
There’s probably a lot more than just that quick Sunday list. There is a lot more that can be done to make programming more RAD, and at the same time more accessible. Speeding up and simplification of common tasks doesn’t necessarily lead to dumbing down of the application capabilities. Separation of concerns, use of logical tiers which can be isolated and swapped in and out again can address all of the “Access in the enterprise nightmares” developers are so quick to recall. I think the biggest hurdle will be getting the professional developers providing solid and safe access to enterprise data: to provide those “cloud” services and then have the tools that let users work with that visually as best suits them.
As to last night’s prank, well it was a child’s toy that made a meow sound when one of the buttons was pressed (unbeknown to me at the time). So at almost 1 AM there I was peering under the house with a torch looking for a lost kitten that never really existed. I sure hope KittyHawk doesn’t turn out to be so elusive as that kitten was ;)