Joe Coder

Apple to gain real ground in dev?

One of Microsoft's lasting areas of dominance has been software development. No one, really, has as advanced an infrastructure, toolset, platform, community or as much momentum. But Apple is wisely recognizing this largely unchallenged market and they've got my attention. First they revealed yesterday that they're now shipping some real powerhouse machines with quad Xeon processors and 16GB of RAM if you want it. Then this: Ruby on Rails will ship with OS X 10.5 (Leopard).

That's a big deal. They're starting to appeal to nerds, not just the snobsquad of thick-frame-glasses-black-turtleneck-kenneth-cole-styled designers.

Microsoft and our community ignore the RoR movement, the clean, simple, lighter, standards compliant, "web 2.0-ish" way of doing things at our peril.

It’s been no secret that Apple is held in very high regard by the Rails community. Every single Rails Core contributer is running on Apple and the vast majority of Rails developers are too. (-emphasis mine, src)

Free copy of WinRAR

Special One Day Offer.

Form Validation Quandary

I'm doing form validation, and it seems I have the following (acceptable*) choices.

1. Straight clean Javascript on the client, double-checked by the server on submit.
2. Ajax talking to the server for the validation logic, reporting back to the client, and of course, backed up by the server.

Issues with option 1:
It's nice because: 90%+ of users will likely get their needs met with the client side checking, very quickly. If it fails, the server logic will keep it from running afoul.

It stinks because: I'll have to write validation logic twice. I'll have to do some additional voodoo to make the client-side reporting of the errors work from the server side.

Issues with option 2:
It's nice because: I only have to write my validation logic once. I can use the same logic, in fact, for the double-check on the server when the page submits.

It stinks because: I have to write the AJAX hooks, and I'll still have to find a way of making the client-side reporting of errors work after the submit.

Maybe it's a wash? Maybe if I go with 2 I'll be able to leverage the AJAX for other benefits.

Additional considerations:
This is another of many moments when I'm faced with the option of just not supporting browsers without Javascript enabled. I've come a long way without making that decision, but it's a hard one to avoid. We know it's not a good practice to do so because it alienates some visitors. But the compromise between creating graceful degradation and "getting something out there" is a tough one. 37signals decided long ago that anyone without Javascript was out of luck. They also have a very sophisticated audience, for the most part. I don't imagine they've lost many clients because of that decision. (I'd love to know the numbers). More importantly, they've released product and are making money.

* The built-in asp.net validation controls are too messy.