Error handling in C code is my typical example of that. It mostly avoids the need for lots of if statements to make sure that you clean up all that you need to and nothing more.
There are other ways to go about it, but in general I'm not convinced they are better.
Sorry, I phrased what I meant poorly. What I really meant is that "going from no version control to CVS brings more benefits than going from CVS to SVN, and that going from CVS to SVN brings more benefits than going from SVN to Git."
This is also a YMMV... I'm also still talking about small centralized teams or even a single person. If you have larger teams things may change; if you have actually decentralized teams, then things almost certainly will change. However, I would say it's definitely true for me and the sorts of projects I'm working on.
The workflow with something like Git or Mercurial is very helpful even if you have a strongly centralized team. You can work on your project on the bus (or at home after you move into a new apartment and don't have internet access for a couple weeks, not this this happened a couple weeks ago to me) without worrying about hoarding changes that you can't commit yet and branching (again useful even if you only have one person) is far less painful than it is with svn. (To be fair I haven't had a chance to try the merge tracking features in 1.5. But at least pre-1.5, the difference was night-and-day.)
Now, that said, I know Subversion really well, and when I tried Git last there were a couple times I managed to get my working directory into semi-corrupt states that I couldn't figure out how to solve. I also think that going from no version control to CVS is a larger step than going from CVS to SVN, and that going from CVS to Subversion is a larger step than going from Subversion to Git, so I am finding it hard to actually use it myself. Besides, the main project I'm working on now has me stuck on CVS ("Crappy Versioning System").
But you think that companies stop caring about a lost $100,000 just because they've made $10,000,000? They don't.
That depends on whether they think they made another $500,000 from people who would have pirated their program but didn't. (They don't have to be right, just think that.)
...and the splitting of content into three separate games...
As opposed to just two with the first one?
I know it's just one, but it's going to be repeated a substantial number of times.
Substantial in relation to the number of sales though? I'd wager almost certainly not. Sure, there are a couple dozen people here on/., but most people aren't going to know, aren't going to care, or (like me) will care but not enough to prevent a purchase (especially if Battle.Net just does matchmaking and local games are still actually played locally, like will probably be the case).
In what sort of crazy FUD-land can you only install Windows once? I've reinstalled Windows with the same license multiple times on several occasions. For XP, that's included many computers with entirely different specs over several years.
Mass Effect is a kind of weird thing for me. On one hand... it's really good. It's got a good story, basically all the dialog is decent at worst, and there are a couple plot revelations that are just extraordinary. There's one that is pretty damn bone-chilling. On the other hand, there are a few things that are a bit aggravating. Most of these are somewhat sort of "technical" issues, and if they were changed basically wouldn't change the game itself at all.
For me, there are a few unfortunate glitches... one one of the missions the ground isn't rendered properly so it's very hard to figure out where to drive without going off of a cliff, and the last half of a mission has lots of sound glitches (e.g. cut scenes without most of the sound tracks, occasional dropped dialog).
I also agree with most of what the Penny Arcade people had to say about it, though I do think the first comic is the least relevant. The first time I played, I had a rather hard time on the harder battles for a while, even on the easiest difficulty setting. Then things clicked and I figured out how things worked, and it got much easier; I'm playing it on hardcore now (the 4th of 5 settings; the 5th isn't unlocked yet), and it's really in the same ballpark as the early parts on casual since the game is harder but I am better.
The PA people didn't make a strip about it, but Tycho discusses it in the third news post in that series: the inventory management system is terrible. It sucks. It's the sort of thing that makes me think that BioWare should take some lessons on playtesting from Valve or something, because I can't imagine that they playtested the current design, properly solicited feedback from the users, and didn't find out how bad it is. The biggest problem is that your inventory can hold 150 items, but there's no internal organization; it's just a list in the order you picked things up. There are nice categories ("pistols", "sniper rifles", "grenade upgrades",...) but they don't show up in the list. You can have two things of the same item that show up in multiple places in the list. A very simple change that would have made this far less frustrating is to set up a hierarchy by genre and a sub-heirarchy by the manufacturer/model. (There are items like "Kessler I", "Kessler II", etc. where a "Kessler II" pistol is strictly better than a "Kessler I" pistol, but incomparable to the "Edge II" pistol. So group all the Kesslers together.) There are other problems too.
Finally, there are just a bunch of misc. annoyances. If you're in a dialog scene, you can't even get to a menu to, say, exit the game. You can't skip cut scenes (even ones that are very repetitive, like going through a mass relay); this goes along with Penny Arcade's elevator strip. There are fights in the game where you have dialog, then a fight, and it will "helpfully" save before the dialog instead of after, so if you die a bunch of times you have to go through the dialog again. At a minute or two a piece just from that, I probably could have saved 10 or 15 minutes earlier today at the end of the mission where you pick up Liara.
I would say don't take an overall negative impression from this review; definitely don't do so based on the volume of text I devoted to complaining. I'm just much better at talking about what I don't like about something than I am at talking about things that I do like. And there is plenty in that latter category. I might summarize it as the following: Mass Effect is a great game when you're done playing it, but only decent while you are actually playing it. This is similar to and inspired slightly by something Tycho wrote: "by the time you are done playing, you remember the emotional topography of the game much more than the technical one". Maybe not worth the $50 launch price or whatever it was, but definitely worth the $20 now.
...Steam's hullabaloo it raises everytime you go offline for five minutes...
I'm hardly a fan of Steam and I think their offline mode is poorly implemented in a few different ways (e.g. a half-updated game is unplayable), but "hullabaloo it raises everytime you go offline for five minutes"? What?
I just moved a few weeks ago, and I didn't have an internet connection for a couple weeks after moving in. A friend and I both finished the Steam-based Mass Effect during that time, running Steam in offline mode.
Didn't have any problems, nor did we get nagged by Steam during that time except at startup.
I was also not a fan of Mostly Harmless; wikipedia says Adams said that he was disappointed with the bleak nature of the book, and would have written the 6th more upbeat; it's quite possible that this would have improved it.
(Incidentally, anyone think that the reviewer missed the perfect opportunity to call the new book "mostly unfunny", or "almost but not quite entirely unlike humor"?)
They don't charge you for the landline if you get DSL? Over here, if you want DSL you have to pay for a landline, and then also pay for DSL on top of that ($30 on top of whatever the landline costs).
It's the same here; my statement was that the combined cost of the land line + DSL was less than cable.
I did a little more investigating and this isn't quite true; it's certainly less than what I was paying for cable at my old address ($45/mth with DSL vs $65/mth with cable), but I had considerably better service with cable than I chose with DSL.
A fairer comparison for the service I have would, in fact, put cable cheaper. (Though for faster service, DSL pulls ahead. 4mb is $50 for DSL and $55 for cable, though with a 6-month $30 rate. 10mb is $60 for DSL, $65 for cable with a 6-month $40 rate.)...I would have to buy a phone (or do they give those away for free with a contract?)
No, they don't. On the other hand, lots of people (me for instance) already have one.
Look, I'm not saying a land line makes sense for you or the OP; just that saying "the ONLY advantage a land line has" is the ability to hook up multiple phones is an awfully narrow view in general.
I have better call quality going cell phone to cell phone than I do from my cell phone to my parents' landline. I'm not sure why their phones are so bad, but they are.
Your experience differs from mine then.
No one forgets to unsilence their landline, because they don't silence them in the first place (if they did, they'd probably forget about it).
Exactly my point. I mean, the fact that you carry your cell phone with you means that you'll have to silence it from time-to-time, sometimes unexpectedly. If you don't bring it with you, then why have a cell in the first place? (Obviously there's a middle ground here that I'm ignoring where you leave your phone behind when you know you'll have to go somewhere where it'll be silenced. I still don't think most people do this.)
I already have a cell phone plan; it would cost me $10 to add another line.
By contrast, I'm not sure how much it would cost me. I suspect that I can't do it with pay-as-you-go, which means a $5-10/mth increase in cell phone cost just to get to a plan before I might even be able to add an extra line. At the same time, I have DSL from a company that (at least said, without me prodding much) that I need to get a land-line with them to have DSL. Which means the land line is... free!
(And it's still cheaper than what I've been paying for cable, though I suspect I could do better if I wanted to bargain with them.)
Besides, if you want something like a landline, VoIP is a better solution in any case.
"Pidgin" is just a fancy word for the low-class broken English that most American blacks speak. Look it up if you don't believe me. So as far as I'm concerned, it never had any credibility in the first place.
What? Way to project your own biases. "Pidgin" languages are any sort of conglomeration languages that develop when you have two peoples that don't have a common language who have to communicate.
In fact, the "low-class broken English that most American blacks speak" (let's even ignore the glaring inaccuracy of that phrase) is really not a pidgin language at all.
I pointed out that with most cell phones, long distance calls are free, he countered with the ONLY advantage a landline has -- you can have two phones on the same line.
That's far from the only advantage. Land lines have better quality (I occasionally have to ask people to repeat themselves on the cell phone because the signal will drop out a little bit), land lines don't have batteries that go dead if you forget to charge them, land lines are cheaper unless you make a bunch of long distance calls, it's basically impossible to lose (even temporarily) your land line phone, you can't leave your land line at work and forget to bring it home, you can't (well are much less likely to) set your land-line to silent and forget to turn it back on and miss calls, you're much less likely to just flat out not hear your phone ring because it's on another floor from you and the ringer is pathetic, you don't have to sprint to the phone's location and hope that you get there fast enough (okay, these last two are sort of covered by "you can install multiple lines", but there are really several benefits to that), and local calls are almost always entirely free with a land line (while with a cell you can use up all your minutes even if people rarely do; if you have a prepaid, pay-as-you-go plan, as I did for a while, then that may be real savings),.
I'm not saying that they are better, but there are a number of advantages.
To be fair, there was a mission in the original SC's (not BW) Terran campaign (around the 4th or 5th) where you have to defend against waves of zerg attacks. In retrospect, it was actually relatively close to this concept, except that the "day" and "night" weren't so well-defined.
So not only is this not new in general, it's not even new to Starcraft.
Yet the user's experience is only little refined because of it.
I don't want to knock that; believe me, I have enough problems with Firefox on Linux because of the lack of separation between the tabs that I can't wait for when Chrome has a decent Linux build. From what it sounds that this guy wants to do, it sounds like he doesn't want behind the scenes changes, he wants to revamp the user experience. (Whether or not this browser will, or will in a good way, we'll have to wait to see.)
That is sheer bullshit spread by software coders too stupid and inexperienced to be able to write good code and test it properly.
To be fair, testing is now how you verify that a program is bug free. (It's basically impossible; because using testing to prove correctness relative to a spec, which is what we're talking about here, would require running a test with every single possible input.)
Basically, put it this way. There are three properties we would like to have of a verification engine:
1. Soundness -- if the verification engine says a program has a property (e.g. "this program will necessarily halt" or "this program will not dereference a null pointer"), then the program actually does have that property.
2. Completeness -- if the verification engine says a program does not have a property, then the program actually does not have that property.
3. Termination -- the verification engine will halt on all inputs.
"All" Rice's theorem says is that you can't have all three properties. If you want, you can have any two of them. If you choose #1 and #2, then if you get an answer from your verification engine, you can trust it.
"So you can't trust the output of any recursive function that distinguishes recursive functions by a non-trivial property."
Yes you can. Suppose that you want a decider that answers questions of the form "does this other program halt". There are several options for this program:
1. If the decider terminates, it will always be correct. (Rice's theorem says there are input programs to the decider for which the decider will not terminate.)
2. The decider will always terminate, but will sometimes say that the input program does halt when, in fact, it does not. However, if the decider says that the program will not halt, it is correct.
3. The decider will always terminate, but will sometimes say that the input program does not halt when, in fact, it necessarily does. However, if the decider says that the program will halt, it is correct.
Rice's theorem does not say that you can't statically determine anything about any program ever.
Incorrect, and a common misconstruation of the halting problem (or really Rice's theorem, which is a short extension of the halting problem).
What Rice's Theorem says is that for any proof technique, there will be some program that confounds it. That is at least one program will cause it to either not terminate or give an incorrect answer.
What we can still do is verify a lot of programs, and in particular, perhaps even most "realistic" programs.
That works out for the IRS, because I'm not sure why you'd want to do comparisons. However, comparisons between different times is exactly what people want to do with cars a lot. "Should I trade in my car that gets 25 mpg for one that gets 35?" I'd wager that this is the second-most used reason behind MPG ratings in the first place (behind comparing two cars that you're thinking about getting now).
And this comparison suddenly becomes harder if the ratings change over time. Now all of a sudden you have to go look up updated numbers (simply remembering "oh, it was rated for around 25 mpg" is insufficient), make sure that your source is up-to-date, etc.
I mean, it's not a big deal, and it's nothing you can't work around... but it is added complexity to the process for what I think is not much benefit, if any. People have a sense right now of what xx MPG means in terms of cost, and if not you can think about what it means for you and your gas prices.
Now that is actually a useful metric! I like it. Of course it would vary based on the cost of gas, the cost of electricity, and so forth. But hey you could make some reasonable assumptions, apply them country-wide, and you could at least meaningfully compare the values even if you had to adjust them for your local conditions to know how much you're really going to spend.
What happens in 5 years though when gas is, say, 50% more expensive than now? Does the EPA adjust the ratings to reflect the new price? If so, then it makes comparing models of now vs 5 years ago hard. When reading a source, you have to think "now, when was it published?" and make the adjustments. If the EPA leaves the numbers the same, it will soon be as abstract of a number as you have now, only it'll be deceptive since it'll be an outright lie.
Long range electric or efficient internal combustion.
According to this article, what is my impression of the current state-of-the-art of realistic electric vehicles, the Tesla Roadster, is capped at 244 miles on a charge.
Are you going to pay for my second car so that I have one that is both super efficient at city driving (as the Volt is) while allowing me to go long trips as well?
Yes, long range electric would be great. So would the Enterprise's transporter.
$$/mile would never ever work; prices of gas and electricity vary way too much by location and time. You need raw unit numbers.
The real problem is reducing it to one or two numbers; you really want more: (gasoline consumption in gallons, electricity consumption in KWh) per mile (city, highway) driving. You probably also want the electricity range in miles, perhaps for both highway and city driving.
Slashdot Mirror
What exactly is that a good solution to?
Error handling in C code is my typical example of that. It mostly avoids the need for lots of if statements to make sure that you clean up all that you need to and nothing more.
There are other ways to go about it, but in general I'm not convinced they are better.
Sorry, I phrased what I meant poorly. What I really meant is that "going from no version control to CVS brings more benefits than going from CVS to SVN, and that going from CVS to SVN brings more benefits than going from SVN to Git."
This is also a YMMV... I'm also still talking about small centralized teams or even a single person. If you have larger teams things may change; if you have actually decentralized teams, then things almost certainly will change. However, I would say it's definitely true for me and the sorts of projects I'm working on.
I almost totally second this.
The workflow with something like Git or Mercurial is very helpful even if you have a strongly centralized team. You can work on your project on the bus (or at home after you move into a new apartment and don't have internet access for a couple weeks, not this this happened a couple weeks ago to me) without worrying about hoarding changes that you can't commit yet and branching (again useful even if you only have one person) is far less painful than it is with svn. (To be fair I haven't had a chance to try the merge tracking features in 1.5. But at least pre-1.5, the difference was night-and-day.)
Now, that said, I know Subversion really well, and when I tried Git last there were a couple times I managed to get my working directory into semi-corrupt states that I couldn't figure out how to solve. I also think that going from no version control to CVS is a larger step than going from CVS to SVN, and that going from CVS to Subversion is a larger step than going from Subversion to Git, so I am finding it hard to actually use it myself. Besides, the main project I'm working on now has me stuck on CVS ("Crappy Versioning System").
But you think that companies stop caring about a lost $100,000 just because they've made $10,000,000? They don't.
That depends on whether they think they made another $500,000 from people who would have pirated their program but didn't. (They don't have to be right, just think that.)
...and the splitting of content into three separate games...
As opposed to just two with the first one?
I know it's just one, but it's going to be repeated a substantial number of times.
Substantial in relation to the number of sales though? I'd wager almost certainly not. Sure, there are a couple dozen people here on /., but most people aren't going to know, aren't going to care, or (like me) will care but not enough to prevent a purchase (especially if Battle.Net just does matchmaking and local games are still actually played locally, like will probably be the case).
...Windows is valid for one install...
In what sort of crazy FUD-land can you only install Windows once? I've reinstalled Windows with the same license multiple times on several occasions. For XP, that's included many computers with entirely different specs over several years.
What did you think of Mass Effect?
Mass Effect is a kind of weird thing for me. On one hand... it's really good. It's got a good story, basically all the dialog is decent at worst, and there are a couple plot revelations that are just extraordinary. There's one that is pretty damn bone-chilling. On the other hand, there are a few things that are a bit aggravating. Most of these are somewhat sort of "technical" issues, and if they were changed basically wouldn't change the game itself at all.
For me, there are a few unfortunate glitches... one one of the missions the ground isn't rendered properly so it's very hard to figure out where to drive without going off of a cliff, and the last half of a mission has lots of sound glitches (e.g. cut scenes without most of the sound tracks, occasional dropped dialog).
I also agree with most of what the Penny Arcade people had to say about it, though I do think the first comic is the least relevant. The first time I played, I had a rather hard time on the harder battles for a while, even on the easiest difficulty setting. Then things clicked and I figured out how things worked, and it got much easier; I'm playing it on hardcore now (the 4th of 5 settings; the 5th isn't unlocked yet), and it's really in the same ballpark as the early parts on casual since the game is harder but I am better.
The PA people didn't make a strip about it, but Tycho discusses it in the third news post in that series: the inventory management system is terrible. It sucks. It's the sort of thing that makes me think that BioWare should take some lessons on playtesting from Valve or something, because I can't imagine that they playtested the current design, properly solicited feedback from the users, and didn't find out how bad it is. The biggest problem is that your inventory can hold 150 items, but there's no internal organization; it's just a list in the order you picked things up. There are nice categories ("pistols", "sniper rifles", "grenade upgrades", ...) but they don't show up in the list. You can have two things of the same item that show up in multiple places in the list. A very simple change that would have made this far less frustrating is to set up a hierarchy by genre and a sub-heirarchy by the manufacturer/model. (There are items like "Kessler I", "Kessler II", etc. where a "Kessler II" pistol is strictly better than a "Kessler I" pistol, but incomparable to the "Edge II" pistol. So group all the Kesslers together.) There are other problems too.
Finally, there are just a bunch of misc. annoyances. If you're in a dialog scene, you can't even get to a menu to, say, exit the game. You can't skip cut scenes (even ones that are very repetitive, like going through a mass relay); this goes along with Penny Arcade's elevator strip. There are fights in the game where you have dialog, then a fight, and it will "helpfully" save before the dialog instead of after, so if you die a bunch of times you have to go through the dialog again. At a minute or two a piece just from that, I probably could have saved 10 or 15 minutes earlier today at the end of the mission where you pick up Liara.
I would say don't take an overall negative impression from this review; definitely don't do so based on the volume of text I devoted to complaining. I'm just much better at talking about what I don't like about something than I am at talking about things that I do like. And there is plenty in that latter category. I might summarize it as the following: Mass Effect is a great game when you're done playing it, but only decent while you are actually playing it. This is similar to and inspired slightly by something Tycho wrote: "by the time you are done playing, you remember the emotional topography of the game much more than the technical one". Maybe not worth the $50 launch price or whatever it was, but definitely worth the $20 now.
What ending did you and your friend take?
...Steam's hullabaloo it raises everytime you go offline for five minutes...
I'm hardly a fan of Steam and I think their offline mode is poorly implemented in a few different ways (e.g. a half-updated game is unplayable), but "hullabaloo it raises everytime you go offline for five minutes"? What?
I just moved a few weeks ago, and I didn't have an internet connection for a couple weeks after moving in. A friend and I both finished the Steam-based Mass Effect during that time, running Steam in offline mode.
Didn't have any problems, nor did we get nagged by Steam during that time except at startup.
I was also not a fan of Mostly Harmless; wikipedia says Adams said that he was disappointed with the bleak nature of the book, and would have written the 6th more upbeat; it's quite possible that this would have improved it.
(Incidentally, anyone think that the reviewer missed the perfect opportunity to call the new book "mostly unfunny", or "almost but not quite entirely unlike humor"?)
They don't charge you for the landline if you get DSL? Over here, if you want DSL you have to pay for a landline, and then also pay for DSL on top of that ($30 on top of whatever the landline costs).
It's the same here; my statement was that the combined cost of the land line + DSL was less than cable.
I did a little more investigating and this isn't quite true; it's certainly less than what I was paying for cable at my old address ($45/mth with DSL vs $65/mth with cable), but I had considerably better service with cable than I chose with DSL.
A fairer comparison for the service I have would, in fact, put cable cheaper. (Though for faster service, DSL pulls ahead. 4mb is $50 for DSL and $55 for cable, though with a 6-month $30 rate. 10mb is $60 for DSL, $65 for cable with a 6-month $40 rate.) ...I would have to buy a phone (or do they give those away for free with a contract?)
No, they don't. On the other hand, lots of people (me for instance) already have one.
Look, I'm not saying a land line makes sense for you or the OP; just that saying "the ONLY advantage a land line has" is the ability to hook up multiple phones is an awfully narrow view in general.
I have better call quality going cell phone to cell phone than I do from my cell phone to my parents' landline. I'm not sure why their phones are so bad, but they are.
Your experience differs from mine then.
No one forgets to unsilence their landline, because they don't silence them in the first place (if they did, they'd probably forget about it).
Exactly my point. I mean, the fact that you carry your cell phone with you means that you'll have to silence it from time-to-time, sometimes unexpectedly. If you don't bring it with you, then why have a cell in the first place? (Obviously there's a middle ground here that I'm ignoring where you leave your phone behind when you know you'll have to go somewhere where it'll be silenced. I still don't think most people do this.)
I already have a cell phone plan; it would cost me $10 to add another line.
By contrast, I'm not sure how much it would cost me. I suspect that I can't do it with pay-as-you-go, which means a $5-10/mth increase in cell phone cost just to get to a plan before I might even be able to add an extra line. At the same time, I have DSL from a company that (at least said, without me prodding much) that I need to get a land-line with them to have DSL. Which means the land line is... free!
(And it's still cheaper than what I've been paying for cable, though I suspect I could do better if I wanted to bargain with them.)
Besides, if you want something like a landline, VoIP is a better solution in any case.
Why?
"Pidgin" is just a fancy word for the low-class broken English that most American blacks speak. Look it up if you don't believe me. So as far as I'm concerned, it never had any credibility in the first place.
What? Way to project your own biases. "Pidgin" languages are any sort of conglomeration languages that develop when you have two peoples that don't have a common language who have to communicate.
In fact, the "low-class broken English that most American blacks speak" (let's even ignore the glaring inaccuracy of that phrase) is really not a pidgin language at all.
I pointed out that with most cell phones, long distance calls are free, he countered with the ONLY advantage a landline has -- you can have two phones on the same line.
That's far from the only advantage. Land lines have better quality (I occasionally have to ask people to repeat themselves on the cell phone because the signal will drop out a little bit), land lines don't have batteries that go dead if you forget to charge them, land lines are cheaper unless you make a bunch of long distance calls, it's basically impossible to lose (even temporarily) your land line phone, you can't leave your land line at work and forget to bring it home, you can't (well are much less likely to) set your land-line to silent and forget to turn it back on and miss calls, you're much less likely to just flat out not hear your phone ring because it's on another floor from you and the ringer is pathetic, you don't have to sprint to the phone's location and hope that you get there fast enough (okay, these last two are sort of covered by "you can install multiple lines", but there are really several benefits to that), and local calls are almost always entirely free with a land line (while with a cell you can use up all your minutes even if people rarely do; if you have a prepaid, pay-as-you-go plan, as I did for a while, then that may be real savings),.
I'm not saying that they are better, but there are a number of advantages.
To be fair, there was a mission in the original SC's (not BW) Terran campaign (around the 4th or 5th) where you have to defend against waves of zerg attacks. In retrospect, it was actually relatively close to this concept, except that the "day" and "night" weren't so well-defined.
So not only is this not new in general, it's not even new to Starcraft.
..its about the changes under the hood!
Yet the user's experience is only little refined because of it.
I don't want to knock that; believe me, I have enough problems with Firefox on Linux because of the lack of separation between the tabs that I can't wait for when Chrome has a decent Linux build. From what it sounds that this guy wants to do, it sounds like he doesn't want behind the scenes changes, he wants to revamp the user experience. (Whether or not this browser will, or will in a good way, we'll have to wait to see.)
That is sheer bullshit spread by software coders too stupid and inexperienced to be able to write good code and test it properly.
To be fair, testing is now how you verify that a program is bug free. (It's basically impossible; because using testing to prove correctness relative to a spec, which is what we're talking about here, would require running a test with every single possible input.)
Basically, put it this way. There are three properties we would like to have of a verification engine:
1. Soundness -- if the verification engine says a program has a property (e.g. "this program will necessarily halt" or "this program will not dereference a null pointer"), then the program actually does have that property.
2. Completeness -- if the verification engine says a program does not have a property, then the program actually does not have that property.
3. Termination -- the verification engine will halt on all inputs.
"All" Rice's theorem says is that you can't have all three properties. If you want, you can have any two of them. If you choose #1 and #2, then if you get an answer from your verification engine, you can trust it.
"So you can't trust the output of any recursive function that distinguishes recursive functions by a non-trivial property."
Yes you can. Suppose that you want a decider that answers questions of the form "does this other program halt". There are several options for this program:
1. If the decider terminates, it will always be correct. (Rice's theorem says there are input programs to the decider for which the decider will not terminate.)
2. The decider will always terminate, but will sometimes say that the input program does halt when, in fact, it does not. However, if the decider says that the program will not halt, it is correct.
3. The decider will always terminate, but will sometimes say that the input program does not halt when, in fact, it necessarily does. However, if the decider says that the program will halt, it is correct.
Rice's theorem does not say that you can't statically determine anything about any program ever.
It's a good thing that the original AC who posted about the task being impossible was right and your arrogance is justified.
Oh wait, he wasn't.
Incorrect, and a common misconstruation of the halting problem (or really Rice's theorem, which is a short extension of the halting problem).
What Rice's Theorem says is that for any proof technique, there will be some program that confounds it. That is at least one program will cause it to either not terminate or give an incorrect answer.
What we can still do is verify a lot of programs, and in particular, perhaps even most "realistic" programs.
That works out for the IRS, because I'm not sure why you'd want to do comparisons. However, comparisons between different times is exactly what people want to do with cars a lot. "Should I trade in my car that gets 25 mpg for one that gets 35?" I'd wager that this is the second-most used reason behind MPG ratings in the first place (behind comparing two cars that you're thinking about getting now).
And this comparison suddenly becomes harder if the ratings change over time. Now all of a sudden you have to go look up updated numbers (simply remembering "oh, it was rated for around 25 mpg" is insufficient), make sure that your source is up-to-date, etc.
I mean, it's not a big deal, and it's nothing you can't work around... but it is added complexity to the process for what I think is not much benefit, if any. People have a sense right now of what xx MPG means in terms of cost, and if not you can think about what it means for you and your gas prices.
Now that is actually a useful metric! I like it. Of course it would vary based on the cost of gas, the cost of electricity, and so forth. But hey you could make some reasonable assumptions, apply them country-wide, and you could at least meaningfully compare the values even if you had to adjust them for your local conditions to know how much you're really going to spend.
What happens in 5 years though when gas is, say, 50% more expensive than now? Does the EPA adjust the ratings to reflect the new price? If so, then it makes comparing models of now vs 5 years ago hard. When reading a source, you have to think "now, when was it published?" and make the adjustments. If the EPA leaves the numbers the same, it will soon be as abstract of a number as you have now, only it'll be deceptive since it'll be an outright lie.
Long range electric or efficient internal combustion.
According to this article, what is my impression of the current state-of-the-art of realistic electric vehicles, the Tesla Roadster, is capped at 244 miles on a charge.
Are you going to pay for my second car so that I have one that is both super efficient at city driving (as the Volt is) while allowing me to go long trips as well?
Yes, long range electric would be great. So would the Enterprise's transporter.
$$/mile would never ever work; prices of gas and electricity vary way too much by location and time. You need raw unit numbers.
The real problem is reducing it to one or two numbers; you really want more: (gasoline consumption in gallons, electricity consumption in KWh) per mile (city, highway) driving. You probably also want the electricity range in miles, perhaps for both highway and city driving.