| |||||||||||||
The Incredible Magical Pantech UML290The Incredible Magical Pantech UML290Posted Jun 23, 2011 17:20 UTC (Thu) by pflugstad (subscriber, #224)Parent article: Quotes of the week
Dan Williams whole blog entry is priceless, and contains what I would have put in a QoTW:
> WMC is just another binary protocol; essentially encoding structs on the wire but with a bunch of stupid at the front and some idiot at the end.
Man, I wish I had a nickle for everytime I'd run into some idiot protocol like this. Makes me think that the real reason why embedded companies don't want to publish specs on their super-duper special proprietary protocols (or drivers, or whatever) is that they don't want to be embarrassed.
The sad part is that I've seen my own company do silly things like, because they think there's some kind proprietary advantage to it. Sigh...
(Log in to post comments)
The Incredible Magical Pantech UML290 Posted Jun 23, 2011 19:19 UTC (Thu) by iabervon (subscriber, #722) [Link]
My suspicion is that the reason embedded companies don't want to publish specs on their proprietary protocols is that they don't actually know what those protocols are. They stick some stuff they have together, and then they tweak it until it the parts talk to each other and don't get confused, and then they leave it. If they were going to release actual specs, they'd have to reverse-engineer their own device to figure out what the total effect of the layers of workarounds is, and they just don't want to do the work. Dan Williams probably understands this device at this point better than any individual at the company.
The Incredible Magical Pantech UML290 Posted Jun 24, 2011 16:16 UTC (Fri) by raven667 (subscriber, #5198) [Link]
Having worked in companies which sell linux-based appliances, the above has been my experience. Some developers try to be more conscientious but don't have the time resource given by management to do a better job and a number of developers just hack and hack and hack at the thing until they get something to work. I could tell some horror stories about temperamental internal software that had idiotic interfaces and failure modes that couldn't be replaced because "well, it worked..sort of and it's already a sunk cost".
Heck, thedailytwf.com is full of stories like that.
The Incredible Magical Pantech UML290 Posted Jun 25, 2011 9:13 UTC (Sat) by linusw (subscriber, #40300) [Link]
When I reverse-engineered the protocol used in the Creative NOMAD and Zen Jukeboxen (prior to the MTP protocol) I had a similar experience. Yes I think I know these protocols better than Creative by now. :-/
|
|||||||||||||