> That's where the delivery of RSS makes sense to me, along with Mail and News. What I > might call "passive delivery". It gives me a list of articles that I can go to or ignore, > all while I'm doing something else.
Umm, I hate to break it to you, but that's pretty much what newsreaders have always done, since before the web existed, let alone RSS. (Okay, not all newsreaders were capable of conveniently operating "while I'm doing something else", but certainly some of them were. And offline newsreaders retrieve the contents of the articles before you make a decision about reading them, but that was an add-on feature, and anyway many newsreaders didn't support it.)
For email, there have been biff utilities even longer -- pretty much since the beginning of time. Since *before* the beginning of time, if you measure time since 1970 in the Unix fashion. If this is what RSS provides, then RSS is just Yet Another Irrelevant Buzzword.
What RSS really provides, IMO, is syndication -- the ability for a third-party source to gather together things from multiple first-party sources and convey them to you, the first party, all together as one feed. This isn't something that's going to change the world, as some over-excited enthusiasts have claimed, but that doesn't mean it can't be useful.
Yes, Microsoft's buzz about what they're doing with RSS is so much hype. Cest la vie.
> He must be using a pretty significantly different version of IE7 than the Beta 2 > preview that I am using, because the tab implementation on my version is pretty terrible.
I suspect he's talking about the same Beta 2, but *comparing* it to something different from what you're comparing to. Specifically, he's probably comparing to earlier iterations of the IE7 tabbed interface, and saying that it's coming along nicely now, in comparison.
And I would agree. The IE7 Beta 2 tabbed interface doesn't have the polish of the one in Firefox, but it is after all a beta, so I'm cutting it some slack for now. It's already nicer than the tabbed interface in the 0.9.6 release of the Mozilla suite, for instance, and probably 0.9.7 and.8 as well. (I could be off by a couple of point releases; it's been a while. I do remember that the nightlies between 0.9.4 and 0.9.5 were when tabs first landed, and they were pretty rough around the edges for a while. By 1.0, they had improved considerably.)
> Because the tabs have to share space with 9 different buttons/menus,
Yeah, I was hoping to be able to relocate those, but in Beta 2 I don't see any way to do it. I am sure they'll fix that eventually, although I wouldn't bet money on its being fixed by the initial release. Maybe. Hopefully.
> I can only have about 3-4 tab names become virtually indistinguishable. In a Firefox > window of the same size, I can have at least 8 tabs showing before the titles begin to > lose meaning.
Ah, yes, I *remember* when I was stuck with a 17" monitor and used 640x480 or 800x600 to avoid the need for squinting. My condolences.
> As far as I can tell, IE7's tab implementation is only useful for people who never > plan to use more than 2-3 tabs at a time.
Or have a larger screen res. I can comfortably get 4 or 5 tabs when running the beta inside an rdesktop window that's rather smaller than the size I usually keep Firefox. If I hooked the Windows system with the IE beta directly to my monitor and cranked up the resolution, I suspect I could get twelve or so tabs. (Granted, that's still not enough. I've currently got eighteen tabs open in Firefox, and it's not unusual to have twice that many, occasionally even more, in which case the HashColoredTabs extension helps. I was pleased to see that IE7 finally supports shortcut icons in the usual places where other browsers put them, not just in the Favourites.)
> And who thought it was a good idea to hide the menu until I press the Alt key?
I am giving them the benefit of the doubt on that one and calling it a beta bug and assuming they will fix it before release. Obviously if they don't that would be a real problem, but in a beta it's not such a very big deal, because it doesn't effect the way pages render, which is IMO the most important value of this beta, letting web devs see how they're stuff's going to come out. I don't think the beta UI is overall polished enough that anyone sane (outside the IE dev team themselves of course (dogfood principle)) is using it as his regular browser all the time. I'm sure there's the usual lunatic fringe doing that, but that's their problem. I'm sure the rough edges on the UI will be smoothed out somewhat before release.
> I've only seen two types of sites that consistently give me security warnings > about ActiveX (SP2). > > 1. MS update pages > 2. Pr0n and hax0r sites trying to install some weird "requirement" that is probably a trojan
Those "Free Online Anti-Virus Scan" websites use it too. (First time I saw one of those it weirded me out, and I didn't think it was for real at first. "How on earth", I thought to myself, "could a website scan _my_ computer for viruses?" Then I saw it was using ActiveX and the light bulb went on. The website could, of course, not directly do any such thing, but software that could be installed by the browser at the behest of the website could do that.)
> So... Exactly why is it I need it again?
That's why they keep deprecating it further and further. Already it gives you scary warnings if your security settings are cranked up; in IE 7 it will be turned off by default. (If I understand correctly, and I hope I do, that means it won't just give scary warnings by default, but actually _not work_, unless the user goes into some settings someplace and flips a switch.) Eventually some future version will not include it at all, but for the sake of backward compatibility that is going to be a while.
The problem is, the security characteristics of corporate intranets are one thing, and the security characteristics of the internet are another thing. Including ActiveX in an intranet tool would have been okay, if that tool weren't marketed as a browser for the world wide web.
> I am too reliant on [features] to ever give up firefox. > They will have to pry it away from my cold dead fingers.
Huh. Interesting mindset. Personally, I'll switch from Firefox the day there's a better browser available that does everything I need it to do and has a nicer interface. Just like I switched away from the Mozilla suite before, and Netscape before that, and NCSA Mosaic before that, and gopher before that.
> (1) George Bush as president of the US? Good idea. > (2) Osama bin Laden as president of the US? Good idea.
> But I realised that to a good proportion of people (myself included) both of > those would seem like bad ideas.
Context can be important in such evaluations. George Bush (either one) was never my first choice for President, but few of the people who would be could ever get a major-party nomination. Even Quayle was ultimately not able to manage that, though he came reasonably close. Relatively speaking, though, I would hope that George Bush (again, either one) would seem like a better idea for the office than option (2) above. I mean, for all that a lot of people disagree with his politics, he's lived in the US pretty much all his life and genuinely thinks that the USA is a pretty good country, as countries go, and that the average US citizen is worth keeping around on the planet. I don't know about you, but those are things _I'd_ sure like for any serious candidate for US President to agree with. (Not that there aren't problems with the US, sure there are, but that's not the point here.)
Still, the Chairman of the Fed made a much more vivid example.
> Besides, it's nice to have Linux booting on as many platforms as possible. One just > never knows when it's going to be useful...
In particular, I'd be quite leary of buying a particular hardware if Knoppix won't run on it. That little gem has saved my bacon more times than I care to recount.
> couldn't Linux be used as an intermediate step to getting Windows running?
You mean by creating a specialized Linux that runs on EFI and then emulate the BIOS while Windows boots, or something? Sort of a custom dual-stage boot loader?
Dunno. Maybe. But there would still be other issues to solve, I suspect.
> Has anybody considered that Microsoft probably has already "ported" Vista to > the MacIntel Developer platform
It's possible, but I find it unlikely. Microsoft does not seem to have the agaility any more to do such things quickly, and there's not adequate motivation to assign any significant team to it.
> It seems to me that the only good reason to pay those bloated prices for Apple > hardware is that you get to run OS/X.
You are not a member of the intended audience. Please accept our apologies for the interruption, and go about your business as if nobody ever mentioned this.
> And there is practically nothing you can do in Linux that I can't do in OS X.
There are a number of things I can do in my Gnome GUI that you can't do in your Carbon/Cocoa GUI, though. (Sure, you could run X11 on top of your Carbon/Cocoa GUI... or even run it non-rootless in a window, if you want, which would let you have non-Apple panels and things... but that wouldn't integrate well with the native GUI.)
Customizing the behavior of the window manager is a good example. For me, though, the real killer is global color settings. On an Apple system, practically every application uses Evil Blinding White Backgrounds that give me a headache if I try to actually, you know, *use* the computer for more than a few minutes at a time. My eyes are more sensitive to light than average, I guess. I require the background color to be dark (my preference is for #294D4A, but anything dark will do in a pinch; white is right out, though). As a corrolary, the text color has to be light; black wouldn't show up very well, and that would lead to eyestrain.
This was always my biggest beef with MacOS Classic, and also with the BeOS. I was hoping that with OS X Apple would sieze the opportunity to fix this oversight and introduce user color preferences that all well-behaved (non-legacy) applications would be expected to observe, but, they didn't.
I could, of course, use OS X and run all X11 applications, but if I wanted to do that, I could just use Darwin, or for that matter Linux or BSD. I think you'll agree that if I'm not going to use Apple's Carbon/Cocoa GUI, then OS X offers few other advantages over the free unices.
I realize this is a complete non-issue for many people, who are perfectly happy staring into #FFFFFF all day long, but to me that is physically painful, so I'd consider this a deal-breaker. Think of it as an accessibility issue, if you will.
> The main thing to get out of all that is that you never get the private keys.. Ever....
Words like "never" are nonsense in the context of security. When you here them, you can be sure you're talking to someone who doesn't understand security, or is trying to hoodwink you.
> And the hash values can only be reset by rebooting.
Or, by making the TPM stuff *think* you're rebooting. That would probably involve (at least) a soldering iron, but in theory it could be done.
However, my money would be on the private key's being compromised in some fashion. For one thing, you're talking about putting copies of it into tens of billions of chips and distributing them with virtually every new PC. With that many copies of it floating around for people to play with, somebody will figure out a way to read the private key off the thing, even if it means destroying some of the chips in the process. The system is brittle because it relies on an unchangeable key. In other words, it fails badly.
> Apple does NOT want OS-X running on non-Apple hardware.
On the whole. They don't care if a couple of extreme geeks after dozens of hours of tinkering manage somehow to get it working. What they don't want is for it to be *easy* to do, and what they *absolutely* don't want is for you to be able to walk into a store and buy a complete set of non-Apple hardware that the manufacturer can guarantee you will work with OS X and all you have to do is pop in the OS X install CD and click Next a bunch of times.
In other words, they don't want it to be at all common or normal or easy to run OS X on non-Apple hardware. Theoretically possible is okay, though.
The more relevant issue here (for Linux on Intel iMacs) is that Linux already has support for Apple's BIOS-substitute, because it has supported Apple hardware since before OS X started shipping. Windows hasn't and doesn't, so getting it to run on Apple hardware will be harder (than doing the same with Linux), even with the Intel CPUs. Basically, Linux already runs on Apple hardware with PPC processors, and it already runs on Intel processors, so getting it to run on Apple hardware with Intel processors is a matter of messing with compilation options and junk, maybe fixing up some make files and whatnot, shouldn't require really any new code.
> I'd never use text-mode to browse the web (except for kicks).
It is, however, a great way to test the graceful degradation of your own sites. It's overkill for most sites, but then, it's also easy to do, and can show up problems that other methods might not.
> 8 bits? Are you mad? If only we had that much ram to waste! Black on white. Two colors: > ONE BIT DISPLAYS!
The eight bits are for storing the *character*, not the color. You'd need sixteen bits if you needed color, but if you don't need color information then eight is enough. Actually we could get by with seven bits per character, but then it wouldn't be byte-aligned, which would have an unacceptable performance penalty, so the extra bit is spare, and can be used for something unnecessary like bold or blinking or foreign letters.
If we didn't need case or punctuation we could go down to six bits, but that would still not be nybble-aligned, so that's only good for compression.
> Osama bin Laden as chairman of the fed? Good idea.
Now, that made me laugh.
You could have made your same point with some garden-variety example, positing bin Laden for a normal (albeit important) position of leadership (e.g., President of the US), wherein the cheif problems have to do with his ideology and background, but instead you chose to suggest him as Chairman of the Federal Reserve, a position with such special requirements that many people who would make fine US Presidents would be terribly inappropriate choices, at the same time setting him against Greenspan, a veritable icon, whose judgement in financial matters has been trusted by people at pretty much all points on the political spectrum for an entire generation. It's the kind of absurd overkill that doesn't just make your point, but is also hilarious.
> I thought that distinction belongs to Sun SPARC boxes running Linux.
I would have guessed that SGI boxes would run higher than SPARC stations, although certainly both are cleanly more expensive than iMacs (or any desktop computer Apple sells for that matter, even if you include one of Apple's big Cinema/Studio Displays).
I feel sorry for people whose favorite languages don't have an equivalent for the CPAN. All general-purpose programming languages *SHOULD* have such a resource.
> You could make the same argument for esoteric software like relational databases.
Relational databases are used in a VERY wide variety of fields, *MUCH* wider than CAD. Banks have relational databases; retail chains these days have relational databases; political parties use them; medium-sized businesses use them to track their customers, their suppliers, their products... medium-sized non-profit organizations use them; most libraries now have relational databases, even *small* libraries; software development firms use them for things like bug tracking; telemarketers use them; spammers are widely believed to use them. You get the idea.
Strong evidence of this can be seen by looking at the commercial products. Microsoft SQL Server is a fairly important part of Microsoft's product line. The same cannot be said for their CAD product. Frankly most of Microsoft's customers (I'm talking direct customers now, not people who buy an OEM PC that happens to come with stuff bundled) are not aware that Microsoft *has* a CAD product.
Autodesk's estimated sales last year (according to Reference USA) came to about $1.2 billion, gross. Oracle does that much in an average month. IBM also makes substantial money from database software.
Another way to see how much more common relational databases are than CAD software is to look at how many book titles there are for SQL versus CAD, or, perhaps more interesting, the Amazon sales rank in books of the first three titles that comes up for each acronym. (SQL: #1848, #8181, #92,737; CAD: #386,655, #96,442, #550,898. If you find a CAD book with an Amazon sales rank in four digits or fewer I'll eat my hat. Incidentally, Excel makes them both look silly, with two out of the three books having a three-digit sales rank, which is starting to get into bestseller territory, i.e., the subject is practically a household word.) This measure alone is unreliable (e.g., Excel also seems to have higher-selling books than MS Word or Windows XP, which is obviously not a reflection of their relative appeal to ordinary users), but when you combine it with the other factors (revenues, number of industries that use it, number of different uses to which it is put,...), it's just one more indicator. The lack of any really successful open-source CAD project that would be comparable to Postgres (or MySQL, or, frankly, even SQLite) would be another indicator.
Basically, relational databases and CAD are just not comparable. People have casual uses for relational databases. People in companies outside the computer industry use them to set up little in-house web apps for things like keeping track of equipment inventory, things that can go down for two days at a time without costing the company anything beyond the tech guy's time to fix it. Heck, I've got a database at home for my personal music collection, with titles and authors and so forth and ratings, which I use to drive the frequency with which songs are played. Granted, I'm a geek, and my mom and dad don't do stuff like that, but the point is that a database can be used for many different things, unlike CAD, which is pretty limited in its applications.
> From the perspective of a home user / small business those may be options
From the perspective of a home user or small business, CAD in general is something you only ever see in movies. Home users just don't sit down and decide, "I think I'll draw up an orthogonic projection of a machined wheel flange shaft today". The closest anyone outside of the relevant fields (drafting, architecture, and so forth) ever gets to this is using a general-purpose vector drawing application (e.g., Illustrator -- or Inkscape; Inkscape is not at this time competitive with Illustrator for a professional, but it's starting to shape up nicely for a casual user, and I suspect that by the time it's as old as Gimp is now it'll probably be seriously nifty; it won't be CAD though, and isn't designed to be).
Open-source CAD software would be comparable to open-source medical records software, or open-source integrated library automation software. There's a sizeable industry that relies on that type of software, and the commercial competition is nothing like as good as standard software with a wider appeal (such as Office), so on the one hand it looks like there's a ready-made niche; but on the other hand, the market is fairly limited compared to the larger market for more standard types of software, and, worse, you can't count on a lot of enthusiastic hobbyists to get involved (even in a beta-testing role). Furthermore, nobody's going to deploy the thing until it's almost up to the point of the current state of the art. The software that would become OpenOffice had users -- real users -- back when it was still owned by StarDivision, because, you know, it was free, and in any event if it screwed up the worst you would lose is the time it takes to type a document. That works for word processing software, because many people use the software casually. It works for a web browser, because users of very early versions (think: Mozilla Seamonkey M12) can use it for their casual browsing and switch to something more fully baked when they're doing something important. That *doesn't* work for something that *only* gets used in mission-critical scenarios, software for which virtually nobody has a casual use.
I'm not saying open-source can't work for software in those categories, but I *am* saying that the dynamics will be very different than they are for more general-purpose software for which many people can find causal uses. Finding a way to make it work is a problem that may only have to be solved once, but as far as I know it has not yet been done (successfully, to the point of producing something even vaguely comparable to the proprietary alternatives in terms of functionality).
> Nope, Linux user. Drag and drop was my first instinct
Uh-huh. How *long* have you been a Linux user? My first instinct would be something like cp file.wav/mnt/ipod
Although, on second thought, once I had checked that that worked I would then probably write a very short Perl script to read my music database (via Class::DBI -- I already have the relevant subclasses because my rating frontend uses them) and transfer all of my highest-rated tracks to the ipod.
Drag and drop? What is this, 1985? Drag-and-drop interfaces are a royal pain. Using the context menu at least wouldn't require moving the mouse halfway across the screen or more.
> Dreamweaver makes it easy to have a local copy to modify, play with, and test changes on.
What on earth do you mean, "Dreamweaver makes it easy" to do this? It was *always* easy to do this. In fact, the only ways to *avoid* having a local copy would be if your editor supported loading and saving remote files as if they were local (e.g., Emacs does this), or having a shell account on the web server and editing the files there directly. These are both rather more complex setups than the standard paradigm of editing the local copy until you're happy with it and then hitting the "upload" button.
> It also makes it easy to sync your local copy to the actual web server copy (uploading > your modified files, and downloading any files modified by others).
Again, any decent file transfer software does this also. In the days of yore the underlying protocol was usually ftp; these days it's usually scp, but that has very little impact on the user interface. (This does rely on keeping the time close to correct (or at least wrong by close to the same amount) on all the systems involved, but since you didn't mention that, I assume Dreamweaver doesn't solve that issue. Timezones aren't a problem, as long as all the systems know what timezone they're in.)
> That is why I was very disappointed to find that NVU does none of these things.
I don't know what NVU is, but standard software has done all of these things since before the web was invented. Every single feature your message lists is something people have taken for granted since before Windows was an operating system.
Slashdot Mirror
> That's where the delivery of RSS makes sense to me, along with Mail and News. What I
> might call "passive delivery". It gives me a list of articles that I can go to or ignore,
> all while I'm doing something else.
Umm, I hate to break it to you, but that's pretty much what newsreaders have always done, since before the web existed, let alone RSS. (Okay, not all newsreaders were capable of conveniently operating "while I'm doing something else", but certainly some of them were. And offline newsreaders retrieve the contents of the articles before you make a decision about reading them, but that was an add-on feature, and anyway many newsreaders didn't support it.)
For email, there have been biff utilities even longer -- pretty much since the beginning of time. Since *before* the beginning of time, if you measure time since 1970 in the Unix fashion. If this is what RSS provides, then RSS is just Yet Another Irrelevant Buzzword.
What RSS really provides, IMO, is syndication -- the ability for a third-party source to gather together things from multiple first-party sources and convey them to you, the first party, all together as one feed. This isn't something that's going to change the world, as some over-excited enthusiasts have claimed, but that doesn't mean it can't be useful.
Yes, Microsoft's buzz about what they're doing with RSS is so much hype. Cest la vie.
> He must be using a pretty significantly different version of IE7 than the Beta 2
.8 as well. (I could be off by a couple of point releases; it's been a while. I do remember that the nightlies between 0.9.4 and 0.9.5 were when tabs first landed, and they were pretty rough around the edges for a while. By 1.0, they had improved considerably.)
> preview that I am using, because the tab implementation on my version is pretty terrible.
I suspect he's talking about the same Beta 2, but *comparing* it to something different from what you're comparing to. Specifically, he's probably comparing to earlier iterations of the IE7 tabbed interface, and saying that it's coming along nicely now, in comparison.
And I would agree. The IE7 Beta 2 tabbed interface doesn't have the polish of the one in Firefox, but it is after all a beta, so I'm cutting it some slack for now. It's already nicer than the tabbed interface in the 0.9.6 release of the Mozilla suite, for instance, and probably 0.9.7 and
> Because the tabs have to share space with 9 different buttons/menus,
Yeah, I was hoping to be able to relocate those, but in Beta 2 I don't see any way to do it. I am sure they'll fix that eventually, although I wouldn't bet money on its being fixed by the initial release. Maybe. Hopefully.
> I can only have about 3-4 tab names become virtually indistinguishable. In a Firefox
> window of the same size, I can have at least 8 tabs showing before the titles begin to
> lose meaning.
Ah, yes, I *remember* when I was stuck with a 17" monitor and used 640x480 or 800x600 to avoid the need for squinting. My condolences.
> As far as I can tell, IE7's tab implementation is only useful for people who never
> plan to use more than 2-3 tabs at a time.
Or have a larger screen res. I can comfortably get 4 or 5 tabs when running the beta inside an rdesktop window that's rather smaller than the size I usually keep Firefox. If I hooked the Windows system with the IE beta directly to my monitor and cranked up the resolution, I suspect I could get twelve or so tabs. (Granted, that's still not enough. I've currently got eighteen tabs open in Firefox, and it's not unusual to have twice that many, occasionally even more, in which case the HashColoredTabs extension helps. I was pleased to see that IE7 finally supports shortcut icons in the usual places where other browsers put them, not just in the Favourites.)
> And who thought it was a good idea to hide the menu until I press the Alt key?
I am giving them the benefit of the doubt on that one and calling it a beta bug and assuming they will fix it before release. Obviously if they don't that would be a real problem, but in a beta it's not such a very big deal, because it doesn't effect the way pages render, which is IMO the most important value of this beta, letting web devs see how they're stuff's going to come out. I don't think the beta UI is overall polished enough that anyone sane (outside the IE dev team themselves of course (dogfood principle)) is using it as his regular browser all the time. I'm sure there's the usual lunatic fringe doing that, but that's their problem. I'm sure the rough edges on the UI will be smoothed out somewhat before release.
> I've only seen two types of sites that consistently give me security warnings
> about ActiveX (SP2).
>
> 1. MS update pages
> 2. Pr0n and hax0r sites trying to install some weird "requirement" that is probably a trojan
Those "Free Online Anti-Virus Scan" websites use it too. (First time I saw one of those it weirded me out, and I didn't think it was for real at first. "How on earth", I thought to myself, "could a website scan _my_ computer for viruses?" Then I saw it was using ActiveX and the light bulb went on. The website could, of course, not directly do any such thing, but software that could be installed by the browser at the behest of the website could do that.)
> So... Exactly why is it I need it again?
That's why they keep deprecating it further and further. Already it gives you scary warnings if your security settings are cranked up; in IE 7 it will be turned off by default. (If I understand correctly, and I hope I do, that means it won't just give scary warnings by default, but actually _not work_, unless the user goes into some settings someplace and flips a switch.) Eventually some future version will not include it at all, but for the sake of backward compatibility that is going to be a while.
The problem is, the security characteristics of corporate intranets are one thing, and the security characteristics of the internet are another thing. Including ActiveX in an intranet tool would have been okay, if that tool weren't marketed as a browser for the world wide web.
> I am too reliant on [features] to ever give up firefox.
> They will have to pry it away from my cold dead fingers.
Huh. Interesting mindset. Personally, I'll switch from Firefox the day there's a better browser available that does everything I need it to do and has a nicer interface. Just like I switched away from the Mozilla suite before, and Netscape before that, and NCSA Mosaic before that, and gopher before that.
> (1) George Bush as president of the US? Good idea.
> (2) Osama bin Laden as president of the US? Good idea.
> But I realised that to a good proportion of people (myself included) both of
> those would seem like bad ideas.
Context can be important in such evaluations. George Bush (either one) was never my first choice for President, but few of the people who would be could ever get a major-party nomination. Even Quayle was ultimately not able to manage that, though he came reasonably close. Relatively speaking, though, I would hope that George Bush (again, either one) would seem like a better idea for the office than option (2) above. I mean, for all that a lot of people disagree with his politics, he's lived in the US pretty much all his life and genuinely thinks that the USA is a pretty good country, as countries go, and that the average US citizen is worth keeping around on the planet. I don't know about you, but those are things _I'd_ sure like for any serious candidate for US President to agree with. (Not that there aren't problems with the US, sure there are, but that's not the point here.)
Still, the Chairman of the Fed made a much more vivid example.
> Besides, it's nice to have Linux booting on as many platforms as possible. One just
> never knows when it's going to be useful...
In particular, I'd be quite leary of buying a particular hardware if Knoppix won't run on it. That little gem has saved my bacon more times than I care to recount.
> couldn't Linux be used as an intermediate step to getting Windows running?
You mean by creating a specialized Linux that runs on EFI and then emulate the BIOS while Windows boots, or something? Sort of a custom dual-stage boot loader?
Dunno. Maybe. But there would still be other issues to solve, I suspect.
> Has anybody considered that Microsoft probably has already "ported" Vista to
> the MacIntel Developer platform
It's possible, but I find it unlikely. Microsoft does not seem to have the agaility any more to do such things quickly, and there's not adequate motivation to assign any significant team to it.
> why?
For the same reason there is textmode Quake.
> It seems to me that the only good reason to pay those bloated prices for Apple
> hardware is that you get to run OS/X.
You are not a member of the intended audience. Please accept our apologies for the interruption, and go about your business as if nobody ever mentioned this.
> And there is practically nothing you can do in Linux that I can't do in OS X.
There are a number of things I can do in my Gnome GUI that you can't do in your Carbon/Cocoa GUI, though. (Sure, you could run X11 on top of your Carbon/Cocoa GUI... or even run it non-rootless in a window, if you want, which would let you have non-Apple panels and things... but that wouldn't integrate well with the native GUI.)
Customizing the behavior of the window manager is a good example. For me, though, the real killer is global color settings. On an Apple system, practically every application uses Evil Blinding White Backgrounds that give me a headache if I try to actually, you know, *use* the computer for more than a few minutes at a time. My eyes are more sensitive to light than average, I guess. I require the background color to be dark (my preference is for #294D4A, but anything dark will do in a pinch; white is right out, though). As a corrolary, the text color has to be light; black wouldn't show up very well, and that would lead to eyestrain.
This was always my biggest beef with MacOS Classic, and also with the BeOS. I was hoping that with OS X Apple would sieze the opportunity to fix this oversight and introduce user color preferences that all well-behaved (non-legacy) applications would be expected to observe, but, they didn't.
I could, of course, use OS X and run all X11 applications, but if I wanted to do that, I could just use Darwin, or for that matter Linux or BSD. I think you'll agree that if I'm not going to use Apple's Carbon/Cocoa GUI, then OS X offers few other advantages over the free unices.
I realize this is a complete non-issue for many people, who are perfectly happy staring into #FFFFFF all day long, but to me that is physically painful, so I'd consider this a deal-breaker. Think of it as an accessibility issue, if you will.
> The main thing to get out of all that is that you never get the private keys.. Ever....
Words like "never" are nonsense in the context of security. When you here them, you can be sure you're talking to someone who doesn't understand security, or is trying to hoodwink you.
> And the hash values can only be reset by rebooting.
Or, by making the TPM stuff *think* you're rebooting. That would probably involve (at least) a soldering iron, but in theory it could be done.
However, my money would be on the private key's being compromised in some fashion. For one thing, you're talking about putting copies of it into tens of billions of chips and distributing them with virtually every new PC. With that many copies of it floating around for people to play with, somebody will figure out a way to read the private key off the thing, even if it means destroying some of the chips in the process. The system is brittle because it relies on an unchangeable key. In other words, it fails badly.
> Apple does NOT want OS-X running on non-Apple hardware.
On the whole. They don't care if a couple of extreme geeks after dozens of hours of tinkering manage somehow to get it working. What they don't want is for it to be *easy* to do, and what they *absolutely* don't want is for you to be able to walk into a store and buy a complete set of non-Apple hardware that the manufacturer can guarantee you will work with OS X and all you have to do is pop in the OS X install CD and click Next a bunch of times.
In other words, they don't want it to be at all common or normal or easy to run OS X on non-Apple hardware. Theoretically possible is okay, though.
The more relevant issue here (for Linux on Intel iMacs) is that Linux already has support for Apple's BIOS-substitute, because it has supported Apple hardware since before OS X started shipping. Windows hasn't and doesn't, so getting it to run on Apple hardware will be harder (than doing the same with Linux), even with the Intel CPUs. Basically, Linux already runs on Apple hardware with PPC processors, and it already runs on Intel processors, so getting it to run on Apple hardware with Intel processors is a matter of messing with compilation options and junk, maybe fixing up some make files and whatnot, shouldn't require really any new code.
> I'd never use text-mode to browse the web (except for kicks).
It is, however, a great way to test the graceful degradation of your own sites. It's overkill for most sites, but then, it's also easy to do, and can show up problems that other methods might not.
> "Electric shaver?! Ha! What are you, a chick?
> You're not hardcore unless you shave with a rock."
How about shaving with duct tape?
> 8 bits? Are you mad? If only we had that much ram to waste! Black on white. Two colors:
> ONE BIT DISPLAYS!
The eight bits are for storing the *character*, not the color. You'd need sixteen bits if you needed color, but if you don't need color information then eight is enough. Actually we could get by with seven bits per character, but then it wouldn't be byte-aligned, which would have an unacceptable performance penalty, so the extra bit is spare, and can be used for something unnecessary like bold or blinking or foreign letters.
If we didn't need case or punctuation we could go down to six bits, but that would still not be nybble-aligned, so that's only good for compression.
> Yeah, but now you can actually play games on the mac!
> Oh, wait... never mind.
Nah, a really good game of NetHack requires *colored* text...
> Osama bin Laden as chairman of the fed? Good idea.
Now, that made me laugh.
You could have made your same point with some garden-variety example, positing bin Laden for a normal (albeit important) position of leadership (e.g., President of the US), wherein the cheif problems have to do with his ideology and background, but instead you chose to suggest him as Chairman of the Federal Reserve, a position with such special requirements that many people who would make fine US Presidents would be terribly inappropriate choices, at the same time setting him against Greenspan, a veritable icon, whose judgement in financial matters has been trusted by people at pretty much all points on the political spectrum for an entire generation. It's the kind of absurd overkill that doesn't just make your point, but is also hilarious.
Good show.
> I thought that distinction belongs to Sun SPARC boxes running Linux.
I would have guessed that SGI boxes would run higher than SPARC stations, although certainly both are cleanly more expensive than iMacs (or any desktop computer Apple sells for that matter, even if you include one of Apple's big Cinema/Studio Displays).
I feel sorry for people whose favorite languages don't have an equivalent for the CPAN. All general-purpose programming languages *SHOULD* have such a resource.
> You could make the same argument for esoteric software like relational databases.
...), it's just one more indicator. The lack of any really successful open-source CAD project that would be comparable to Postgres (or MySQL, or, frankly, even SQLite) would be another indicator.
Relational databases are used in a VERY wide variety of fields, *MUCH* wider than CAD. Banks have relational databases; retail chains these days have relational databases; political parties use them; medium-sized businesses use them to track their customers, their suppliers, their products... medium-sized non-profit organizations use them; most libraries now have relational databases, even *small* libraries; software development firms use them for things like bug tracking; telemarketers use them; spammers are widely believed to use them. You get the idea.
Strong evidence of this can be seen by looking at the commercial products. Microsoft SQL Server is a fairly important part of Microsoft's product line. The same cannot be said for their CAD product. Frankly most of Microsoft's customers (I'm talking direct customers now, not people who buy an OEM PC that happens to come with stuff bundled) are not aware that Microsoft *has* a CAD product.
Autodesk's estimated sales last year (according to Reference USA) came to about $1.2 billion, gross. Oracle does that much in an average month. IBM also makes substantial money from database software.
Another way to see how much more common relational databases are than CAD software is to look at how many book titles there are for SQL versus CAD, or, perhaps more interesting, the Amazon sales rank in books of the first three titles that comes up for each acronym. (SQL: #1848, #8181, #92,737; CAD: #386,655, #96,442, #550,898. If you find a CAD book with an Amazon sales rank in four digits or fewer I'll eat my hat. Incidentally, Excel makes them both look silly, with two out of the three books having a three-digit sales rank, which is starting to get into bestseller territory, i.e., the subject is practically a household word.) This measure alone is unreliable (e.g., Excel also seems to have higher-selling books than MS Word or Windows XP, which is obviously not a reflection of their relative appeal to ordinary users), but when you combine it with the other factors (revenues, number of industries that use it, number of different uses to which it is put,
Basically, relational databases and CAD are just not comparable. People have casual uses for relational databases. People in companies outside the computer industry use them to set up little in-house web apps for things like keeping track of equipment inventory, things that can go down for two days at a time without costing the company anything beyond the tech guy's time to fix it. Heck, I've got a database at home for my personal music collection, with titles and authors and so forth and ratings, which I use to drive the frequency with which songs are played. Granted, I'm a geek, and my mom and dad don't do stuff like that, but the point is that a database can be used for many different things, unlike CAD, which is pretty limited in its applications.
> From the perspective of a home user / small business those may be options
From the perspective of a home user or small business, CAD in general is something you only ever see in movies. Home users just don't sit down and decide, "I think I'll draw up an orthogonic projection of a machined wheel flange shaft today". The closest anyone outside of the relevant fields (drafting, architecture, and so forth) ever gets to this is using a general-purpose vector drawing application (e.g., Illustrator -- or Inkscape; Inkscape is not at this time competitive with Illustrator for a professional, but it's starting to shape up nicely for a casual user, and I suspect that by the time it's as old as Gimp is now it'll probably be seriously nifty; it won't be CAD though, and isn't designed to be).
Open-source CAD software would be comparable to open-source medical records software, or open-source integrated library automation software. There's a sizeable industry that relies on that type of software, and the commercial competition is nothing like as good as standard software with a wider appeal (such as Office), so on the one hand it looks like there's a ready-made niche; but on the other hand, the market is fairly limited compared to the larger market for more standard types of software, and, worse, you can't count on a lot of enthusiastic hobbyists to get involved (even in a beta-testing role). Furthermore, nobody's going to deploy the thing until it's almost up to the point of the current state of the art. The software that would become OpenOffice had users -- real users -- back when it was still owned by StarDivision, because, you know, it was free, and in any event if it screwed up the worst you would lose is the time it takes to type a document. That works for word processing software, because many people use the software casually. It works for a web browser, because users of very early versions (think: Mozilla Seamonkey M12) can use it for their casual browsing and switch to something more fully baked when they're doing something important. That *doesn't* work for something that *only* gets used in mission-critical scenarios, software for which virtually nobody has a casual use.
I'm not saying open-source can't work for software in those categories, but I *am* saying that the dynamics will be very different than they are for more general-purpose software for which many people can find causal uses. Finding a way to make it work is a problem that may only have to be solved once, but as far as I know it has not yet been done (successfully, to the point of producing something even vaguely comparable to the proprietary alternatives in terms of functionality).
> Nope, Linux user. Drag and drop was my first instinct
/mnt/ipod
Uh-huh. How *long* have you been a Linux user? My first instinct would be something like
cp file.wav
Although, on second thought, once I had checked that that worked I would then probably write a very short Perl script to read my music database (via Class::DBI -- I already have the relevant subclasses because my rating frontend uses them) and transfer all of my highest-rated tracks to the ipod.
Drag and drop? What is this, 1985? Drag-and-drop interfaces are a royal pain. Using the context menu at least wouldn't require moving the mouse halfway across the screen or more.
> Dreamweaver makes it easy to have a local copy to modify, play with, and test changes on.
What on earth do you mean, "Dreamweaver makes it easy" to do this? It was *always* easy to do this. In fact, the only ways to *avoid* having a local copy would be if your editor supported loading and saving remote files as if they were local (e.g., Emacs does this), or having a shell account on the web server and editing the files there directly. These are both rather more complex setups than the standard paradigm of editing the local copy until you're happy with it and then hitting the "upload" button.
> It also makes it easy to sync your local copy to the actual web server copy (uploading
> your modified files, and downloading any files modified by others).
Again, any decent file transfer software does this also. In the days of yore the underlying protocol was usually ftp; these days it's usually scp, but that has very little impact on the user interface. (This does rely on keeping the time close to correct (or at least wrong by close to the same amount) on all the systems involved, but since you didn't mention that, I assume Dreamweaver doesn't solve that issue. Timezones aren't a problem, as long as all the systems know what timezone they're in.)
> That is why I was very disappointed to find that NVU does none of these things.
I don't know what NVU is, but standard software has done all of these things since before the web was invented. Every single feature your message lists is something people have taken for granted since before Windows was an operating system.