Well, I think the proof of this is that DISH DVRs do pretty much the same exact sorts of tasks that DTV DVRs do, but they have historically always done them much snappier, with much fewer delays.westview said:
I am not a cheerleader (nor a fan, really) of DISH; generally speaking I think DTV is a better company, is more competent technically, and treats customers better. But DISH excels in at least this one area, which proves that if you write the code correctly in the first place, sluggishness is not an issue.
But how to fix it is a large question. There are three ways to go:
1) Faster CPUs/larger cache/more RAM. IOW, build a bigger boat.
2) Tweak the OS surgically. It seems that this is what DTV is trying to do, bless their little hearts. A valiant effort, but probably mostly ineffectual, and a case of too little, too late.
3) Rewrite the OS. That would be very effective, and who knows, there may be a black project somewhere in the bowels of DTV busily trying to do this very thing. Its expensive, time-consuming, a moving target of changing complexity, and would have a lot of fallout in the growing pains associated with that.
2) Tweak the OS surgically. It seems that this is what DTV is trying to do, bless their little hearts. A valiant effort, but probably mostly ineffectual, and a case of too little, too late.
3) Rewrite the OS. That would be very effective, and who knows, there may be a black project somewhere in the bowels of DTV busily trying to do this very thing. Its expensive, time-consuming, a moving target of changing complexity, and would have a lot of fallout in the growing pains associated with that.
I think #1 and #2 are a given. I hold little hope for #3. This is the same reason that Windows 7 (and probably Windows 8, although it will contain a lot of new code) is really just an ongoing patch of the original x86 code. XP has hundreds of thousands of lines of original code that are patched around (Windows 7 cleared a lot of that away). And until they start over from scratch, Windows will always be hampered to some extent by its legacy, 30-year-old code.
It is not a small undertaking to try to rewrite from the ground up. The rewrite of Tivo OS a few years ago did not go well. Up until then Tivo was rock-solid, and ever since it has been as flaky as any garden-variety cable DVR. And that was not a full rewrite, although a significantly invasive one. They solved the problems of sluggishness (by the standards of 2005, anyway) and added modern features like folders, but all at a price I think they were not prepared to pay. Buh-bye, Tivo, BTW. We miss you.
But it can be done, and done successfully. The best example that comes to mind is when Apple completely rewrote the Mac OS in 2000. That was not truly from the ground up; it used proven elements such as Free BSD unix, the Mach kernel, and elements from the NeXt OS. But it was implemented mostly seamlessly, and was a remarkable accomplishment. The resurgence of Apple from a company 90 days away from Chapter 11 in the 90's to becoming the world's highest-valued company in 2012 happened for a lot of different reasons, but the one that seems to have gone stealth and never really gets the respect it deserves for the weight it carries is how elegant and beautiful, and near-perfect the Mac OS has been since then. Hire those guys.
Fingers crossed...
