That is the point and the problem. There are a lot of them. The app has to deal with the best common denominator. Not neessarily the lowest, but it has to act in a way that will not be preceived as cryptic by the users.
Lots of user testing will eventually solve this in any application, but the app may never get the chance if it starts off "too hard" and the developers are not willing to fix the problems that are found. Too often the answer will be, "Why change it? *We* know how it works."
Never assume the users will be dying to figure out the guts of your app. If they can't see how they will be getting results in a "reasonable" time, they will be looking to invest that time somewhere else.
Assume "newbie on deadline", then complicate progrm flow as needed from there.
See, you ran into the problem most small app programmers just never seem to consider fully: The users don't care about what you care about. The users don't think the way you do. The users don't act the way you expect them to.
Every individual user will have their own take on "how it should be done". If your app doesn't take that into consideration, it will be dropped as "too hard to use" no matter how hard you worked on it or what cool functionality it gives.
The interface design, GUI or CLI, needs to have the users' point of view firmly in mind or adoption will be low.
The article is well written and conversational for the layman. Great. But he doesn't really go into the one great unknown area oput there - HDTV.
What are the best HDTV capture cards, for Over the Air or for backside-of-the-cable/satelite-box? The article only touches on this, but it will be of greater concern for the home enthusiast/hacker in the next two years.
And by the way, what packages support this? MythTV, Freevo, etc.
Farmers are irrelevant. Here in the State of Bulemia they and thier "food" are only a passing fad, coming and going from time to time, but always being run off.
In the Free State of Anorexia, our neighbors to the south, they and their noxious byproducts are not tolerated at all.
Unfortunately there is a point here. The non-traceability of the Open Source process leaves any given product open to contamination from copyrighted/patented IP. Most projects don't have tight checking of who did what, and they definitely don't know where the contributor got the input. That is an invitation for trouble. Worse, a project could have an "IP bomb" placed inside it by an agent of a less than scrupulous SCO, er... proprietary company that wants to stir up trouble later.
On the good side, it is a problem that is easily fixed. Traceability of the code base back to the contributor can be implemented, but it means some sort of centralized repository AND use of good tracking tools. IMO, no major distribution, and definitely no kernel, should leave the foundry without knowing who touched it.
The Eugenics Wars DID start in 1993. And we know the East Asian alliance will be landing in North America by 2017 at the latest. The amazing thing the intel folks at WestCentCon can't figure out is how Roddenberry got it all so close from back in the 60s?
But enough of this, lets just adjust your VR probe and put you back where you belong...
[link mod US-World-Dom-05]
[run]...
Farscape's Peacekeepr Wars deserves very high praise. Every scene looked like it would have been a whole episode of the next season. They kept it good, kept it tight, kept it moving and got off the stage with the least amount of fluffing around.
Perfection would be if the new Star Wars was opening the same day. Killing off one flawed show while another cursed franchise opens on Friday the 13th would be ideal.
Sadly, Lucas foresaw that and is opening on the 19th instead. All the Enterprise refugees will be out of primary shock by then and ready to absorb anything he flings at them.
This is NOT about making the hardware Open Source, it is about making the hardware minimally (or maximally) compatible with Open Source software.
The manufacturer needs to know what a driver builder will need to know to keep the drivers current. He ALREADY knows they WANT to know everything. But as a manufacturer trying to control his product, he may not want people writing software hacks to do any damn thing they please with the hardware. So he DOES NOT want to just release every detail imaginable. Just enough to implement the functions the device is supposed to do. Interface Control Documentation. You stay on your side, I stay on mine.
If the internal scheme is 5 bit instructions on 9 bit boundaries, you need to know that. But you don't need to know the command sequence that spits sparks out the power supply fan.
Mr Nancy was in "American Gods" as one of the semi-protagonist characters. He's portrayed as a down on his luck older Black man, but you can see he's cagey. It is made clear he is Anansi in retirement.
Feeding the troll just his once...
Looking at an AMD product name tells me real performance information, i.e. a 3200 is nominally 6.66% faster than a 3000. Looking at Intel, a P4 540 is faster than a 538, but how much?
AMD could make it clearer in their FX and Opeteron lines, but once you know the performance of one, you can estimate the others.
but it isn't like you cared about the real answer anyway.
The trouble with Intel products is trying to know what they're good for. In the rush to be something for everybody, they have saturated their own market and it is very difficult to tell the real benefits of one processor over another.
Why would I want a 540 over a 530 over a 520 ? I assume price, but is there something else? And the same goes for the Centrino and other lines. Not that they are useless, but there is no clear statement, "you need an M processor for that problem".
AMD has been better about this, trying to clearly differentiate AMD 64 from the FX line. And they named Opteron a whole new name, even though the 1xx is very similar to the FX line. Good move on all counts. And when AMD gets their version of hyperthreading working, it will be better yet.
Slashdot Mirror
That is the point and the problem. There are a lot of them. The app has to deal with the best common denominator. Not neessarily the lowest, but it has to act in a way that will not be preceived as cryptic by the users.
Lots of user testing will eventually solve this in any application, but the app may never get the chance if it starts off "too hard" and the developers are not willing to fix the problems that are found. Too often the answer will be, "Why change it? *We* know how it works."
Yep.
Never assume the users will be dying to figure out the guts of your app. If they can't see how they will be getting results in a "reasonable" time, they will be looking to invest that time somewhere else.
Assume "newbie on deadline", then complicate progrm flow as needed from there.
See, you ran into the problem most small app programmers just never seem to consider fully:
The users don't care about what you care about.
The users don't think the way you do.
The users don't act the way you expect them to.
Every individual user will have their own take on "how it should be done". If your app doesn't take that into consideration, it will be dropped as "too hard to use" no matter how hard you worked on it or what cool functionality it gives.
The interface design, GUI or CLI, needs to have the users' point of view firmly in mind or adoption will be low.
The article is well written and conversational for the layman. Great. But he doesn't really go into the one great unknown area oput there - HDTV.
What are the best HDTV capture cards, for Over the Air or for backside-of-the-cable/satelite-box? The article only touches on this, but it will be of greater concern for the home enthusiast/hacker in the next two years.
And by the way, what packages support this? MythTV, Freevo, etc.
Farmers are irrelevant. Here in the State of Bulemia they and thier "food" are only a passing fad, coming and going from time to time, but always being run off.
In the Free State of Anorexia, our neighbors to the south, they and their noxious byproducts are not tolerated at all.
Unfortunately there is a point here. The non-traceability of the Open Source process leaves any given product open to contamination from copyrighted/patented IP. Most projects don't have tight checking of who did what, and they definitely don't know where the contributor got the input. That is an invitation for trouble. Worse, a project could have an "IP bomb" placed inside it by an agent of a less than scrupulous SCO, er... proprietary company that wants to stir up trouble later.
On the good side, it is a problem that is easily fixed. Traceability of the code base back to the contributor can be implemented, but it means some sort of centralized repository AND use of good tracking tools. IMO, no major distribution, and definitely no kernel, should leave the foundry without knowing who touched it.
The Eugenics Wars DID start in 1993. And we know the East Asian alliance will be landing in North America by 2017 at the latest. The amazing thing the intel folks at WestCentCon can't figure out is how Roddenberry got it all so close from back in the 60s? But enough of this, lets just adjust your VR probe and put you back where you belong... [link mod US-World-Dom-05] [run] ...
Farscape's Peacekeepr Wars deserves very high praise. Every scene looked like it would have been a whole episode of the next season. They kept it good, kept it tight, kept it moving and got off the stage with the least amount of fluffing around.
Season 5 would have been great.
Perfection would be if the new Star Wars was opening the same day. Killing off one flawed show while another cursed franchise opens on Friday the 13th would be ideal.
Sadly, Lucas foresaw that and is opening on the 19th instead. All the Enterprise refugees will be out of primary shock by then and ready to absorb anything he flings at them.
Wrong.
This is NOT about making the hardware Open Source, it is about making the hardware minimally (or maximally) compatible with Open Source software.
The manufacturer needs to know what a driver builder will need to know to keep the drivers current. He ALREADY knows they WANT to know everything. But as a manufacturer trying to control his product, he may not want people writing software hacks to do any damn thing they please with the hardware. So he DOES NOT want to just release every detail imaginable. Just enough to implement the functions the device is supposed to do. Interface Control Documentation. You stay on your side, I stay on mine.
If the internal scheme is 5 bit instructions on 9 bit boundaries, you need to know that. But you don't need to know the command sequence that spits sparks out the power supply fan.
Mr Nancy was in "American Gods" as one of the semi-protagonist characters. He's portrayed as a down on his luck older Black man, but you can see he's cagey. It is made clear he is Anansi in retirement.
Feeding the troll just his once... Looking at an AMD product name tells me real performance information, i.e. a 3200 is nominally 6.66% faster than a 3000. Looking at Intel, a P4 540 is faster than a 538, but how much? AMD could make it clearer in their FX and Opeteron lines, but once you know the performance of one, you can estimate the others. but it isn't like you cared about the real answer anyway.
OK, nobody's perfect... Intel saturates their own product lines, AMD murders theirs.\
The trouble with Intel products is trying to know what they're good for. In the rush to be something for everybody, they have saturated their own market and it is very difficult to tell the real benefits of one processor over another.
Why would I want a 540 over a 530 over a 520 ? I assume price, but is there something else? And the same goes for the Centrino and other lines. Not that they are useless, but there is no clear statement, "you need an M processor for that problem".
AMD has been better about this, trying to clearly differentiate AMD 64 from the FX line. And they named Opteron a whole new name, even though the 1xx is very similar to the FX line. Good move on all counts. And when AMD gets their version of hyperthreading working, it will be better yet.