DJ Chitlenz There Is No Cool

Ok, so I haven’t written a whole lot about Touch Medical, since we’ve been creating it over the last year or so, but it’s finally coming together for a v1 FDA filing, and it’s really forward thinking and new. Some observations about VS2010 and developing for WPF…

1 – Infragistics is THE SHIT. It gave us a way to wrap a lot of “nice to haves” into what has traditionally been a cold clinical interface.

2 – VS2010, despite SP1, still isn’t completely baked yet. It’s still way faster and easier to develop for forms, and XAML doesn’t debug very well AT ALL. I think the team thinks it was worth it to be forward thinking and adopt a new language, but MS seriously needs to re-evaluate marketing XAML. It does NOT reduce work, it actually complicated things exponentially.

3 – Design patterns are a trainwreck under WPF-MVVM. We set out to use prism and fork our apps to phone7/Silverlight/WPF under one codebase. Not gonna happen. When design patterns intersect with reality you end up with tons of interops to deal with old vendor code. A good for instance is Screens, which in a multiple monitor system is required, but it nowhere to be found in XAML to date. This sucks, and is especially painful in medicine where multiple fullscreen footprints are NORMAL. MVVM is great until it isn’t (one of many noble goals that completely ignore the practical world), and while we enforced it effectively in the end, it took WAY longer than it should have and is WAY overcomplicated.

4 – Microsoft… WTF TFS? Man, it took a week to install, and has proven to be a pain in the ass all the way along. They should have kept sourcesafe separate, and not added the Sharepoint baggage, it used to just work…

Those are just some quick thoughts as we wrap up QA of version 1. The upside of pushing forward and dealing with VS2010 and WPF as they are is that we got way ahead of the curve (as long as WPF doesn’t disappear, and I think XAMl is now here to stay).

Anyway, about another week, and I’ll have a second FDA filing under my belt (still not getting rich though… maybe this time /crosses fingers).

-chitlenz

We couldn’t get Dallas to let go of the old STAIR code, which kind of sucks in the end, but at the same time I DID get my old dev. team back, meaning Gerhard and Jory are along for the ride again. It works out for the best anyway I guess, since the STAIR code won’t fit into the HL7 mold very well anyway, and it was going to be a hassle to cram it into a hole that didn’t fit (no, not your mom, sicko!).

Anyway, so Gerhard started last week, and it’s down to SDLC filings over the next few days to finish up the Touch Medical design. Agile is good. When we wrote STAIR, it was the infamous “one off” approach, which kind of makes it amazing we got anywhere at all (in the end, it turned out to be an exercise in pure will to haul the project across the finish line). The fucked up thing was that we had to do all the Agile/SDLC crap to file FDA, but because we wrote to describe a process we had already done, we saw very little benefit from the method. This time around, the 4 month drag of trying to buy STAIR gave me a good long time to review all the questions around our design and get them nailed down. I bet we save years with a roadmap…

C# still rocks, and with the sudden emergence of mobile devices this year (such as the ipad), it’s looking like the WPF/Silverlight evolution of the Visual Studio framework is a winner. With C# at the back of WPF, we are able to separate the design and code roles out giving us an easy way to allocate people’s talents. It’s going to be 4 apps, 5 utilities, a web portal, and 4 servers. We have 4 guys. The timer starts for real Nov 1. Another software project trip… here we go…

I had to eject the STAIR website (well … duh!), and it fucked up the links in my blog something awful. I refreshed the wordpress theme, and patched it all up to date, but some of the older links may be broken for awhile, so bear with me as I start to clean up the mess.

It looks like STAIR got final notification of FDA conformance in early April.

Just for the records, I had to WRITE most of that shit, probably 1000 or the 1300 pages submitted, and boy wasn’t it a ton of fun with essentially no reward. This is why I didn’t post many blog entries during that time (writers block?).

Anyway, I’m back on the path to do it again for touchpacs, so I’m either the dumbest man alive, or the most insane… time will tell I spoze.

–chitlenz

Device Classification Name system, image processing, radiological
510(k) Number K100236
Device Name STAIR SYSTEMS PACS & DICOM VIEWER
Applicant
STAIR SYSTEMS, INC.
747 cedard hill drive
allentown, PA 18109
Contact james w monroe
Regulation Number 892.2050
Classification Product Code
LLZ
Date Received 02/02/2010
Decision Date 03/29/2010
Decision substantially equivalent (SE)
Classification Advisory Committee Radiology
Review Advisory Committee Radiology
Statement/Summary/Purged Status Summary only
summary summary
Type Traditional
Reviewed by Third Party No
Expedited Review No

This may seem like an easy feat for most folks, to explain their jobs, but when you work on future products with markets that don’t exist or in which the technology of the software is just emerging it gets … difficult.

Here are some early screenshots of the 3 montor Northstar PACS system we are shortly finalizing for FDA approval.  Enjoy =)

Kid Rock – Audio Test