WorldCom has close to $30B of debt. The interest payments alone are huge.
Sort of. If they court lets WCOM go into chap 11 the debt is all susspended, along with any pre-chap 11 bills of any kind. To re-emerge from chap-11 they have to show that they are no longer losing money each month, and that they have a "repayment" plan worked out with their creditors. The repayment plan is normally some variant of "screw you, for ever $1 we owed you you get $0.03 and one share of stock...ha ha ha!".
In other words chap 11 basically lets you strip off all the debt (or a huge percentage of it). It also lets you out of any long term contracts (to buy or sell!) that you didn't like, makes it easier to fire people. Oh, and they can "un lease" a building by taking all their stuff out of it and saying "no, were are not paying" to the landlord.
So "too much debt" is no longer WCOMs problem. It is the problem of banks that loaned money, and supplyers that sent goods on purchase orders.
Now I have a theory that the little telco's going bankrupt forced prices down (the little telcos suddonly had "free" infrastructure!) which screwed the medim size ones...and so on. Not just the bankruptcy of corse, it was mostly the drop of the "dot coms", but still I think the bankruptcy laws do have a bit of a chain reaction type thing.
Assuming they are selling a service at a cost that at least covers current costs (ignoring build out a deployment cost!) then either they will keep running it in chap 11 (they can choose to break contracts without peanilty...so they may decide to stop offering a service if they are not making money on it! they can also decide to renogiate any long term contacts they wish, even if they are profitable)....er where was I? Ah, either they will run it in chap 11 and keep running when/if they re-emerge from bankruptcy, or they will sell it off with the infastructure to some company that wants to keep running it.
If the service is not profitable they may try to raise the price, or sell it off, or discontinue it. This is pretty much the golden opertunity for that sort of thing, so if you are getting a "great deal" on something from them, that might not last.
It means if they do let you go they only have to pay out about two weeks of vacation tops as severence. There are also limits on expense reimbursment. On the other hand you paycheck is about as protected as it gets. Unless of corse they lay you off.
It also means they are (I think) less stringent requirments for the layoffs and such. It also means they are prohibited by law from paying off any pre-chap-11 debt, and nobody can deny then continued service on contracts so long as they pay the new bills. Which makes it a lot eaiser to stay in busniess.
Nope, that is just syntax to parse. Complex messy syntax, but still just syntax. Once you parse it, you treat it like a function call.
Templates are more then syntax because use of a template creates a new type on the fly, so your type tracking system has to be able to handle making types when it sees templates.
On the plus side, at least you are not using ObjectaveC which lets you pass messages to objects of a totally opaque type. At least with Java/C++ you always know (er, except with dynamicly loaded objects, which you say are normally rare) that a variable X is of type T, or any descendent of T. With ObjC what you frequently know about variable X is it is "some sort of object or other, maybe that someone wrote code for...":-)
There's always someone who has to take the overkill scenario
Oh, I didn't mean to imply overkill so much as if you have a hard problem, sometimes it is easier to solve part of it, and then ask for help, then start off with none solved and ask for help.
I don't know that this is such a case, but it could be!
(personally, I'm less interested in the audit then the function call tree the auditor has to build, and the variable reference tree it should build as well....er, has too if you care about constructor/destructor calls)
Yes, this does make the logic behind dependancy analysis very difficult. My feeling was that, in a given system (which is assumed to be complete) a method cannot be considered audited until all descendants which override it are audited.
That is an interesting approach, but it means your audited libs are not if they expose any virtual functions. Maybe not a big deal. On large projects it means maybe I can't audit all of my code until you audit some of yours, which can be a big deal. It also makes some forward progresses very hard to see.
A long long time ago I wrote a bunch of code analysis tools for APL2 code. For a while I kept getting tempted to re-write them for C. When I learned C++, I just plain gave up. It is so much harder to figure out what function is being called when you just see t->foo()!
I'm quite sure that only the parsing of the Java is easier then C++, not the actual analysis. Er, except maybe for templates. Drat.
our comments and observations show that it would be valuable to gather a team to approach a problem like this; many heads are better than one (usually).
Sometimes. Other times the many-heads produce a really complex set of goals, and the project flounders. A single head could produce a realistically small set of goal, write some code to half do the job, and then kick it on out where 50,000 people go "that's stupid, I could do better", and 3 people submit patches to make stuff better.
However It bothers me as to why it's is a sepereate project. Subversion or CVS v2.0 or CVS (Subversion) v2.0? I don't see why a completely new project is required.
Pretty much all new code. New data structures. Slightly different view of the problem. Totally different network protocalls. And probbably zero interoperability.
Sounds like a good time to change the name. Even if many of the devlopers are the same.
A class cannot be considered audited until all members are audited, and a file requires all contents to be audited.
Implementing a system like this for a languge like Java would not be tremendously difficult; but C++ (especially) is a very complex language to parse and analyse in this manner.
Java can let you load a class by name and call it. The name can be dynamically constructed. So any function that does this is unauitable. (C++ can normally also do this, but there is no standard way)
In both languages any virtual (or non final) member function might be a call to a never seen before class (if, for example you are a libarary).
I don't think C++ is any harder to analyse in this way, both are quite difficult, and full of big nasty traps. C++ is quite a bit harder to parse though.
Instead of whining about Apple's lack of OS X push, i'd like to hear what that rep thinks Apple could do better
Well, they did actually charge people to go from OS9 to OSX...I think. And to go from 10.0 to 10.1 cost $20 if you coulnd't get to a local store (where they were authorised to give out upgrade CDs for free, and burn new ones if the supply ran out even). Roumor has it the 10.2 upgrade won't be free either. All of those things have likely slowed down adoption a little.
However I think what really has them dragging is not forcing people to upgrade wheather the OS is ready or not. I mean MS is doing whatever they can to force people to pay extra to get their old OSes, and here Apple is handing OS9 out for free on new machines. Even having it pre-installed! Why then even go to the effort of making it work on systems designed after OSX was brought out! Heck Apple even made the iPod work with the old OS9 iTunes. The only new software of any signifigance Apple brought out that is OSX only is iPhoto! (er, iDVD maybe too?)
Clearly Apple isn't doing everything they can to get people to go to OSX. Of corse other then (possably) the charging for money part, I think they are doing everything reasonable to get people to switch.
So perhaps one can blame IBM for having started the ball rolling on the idea of strong control of "intellectual property" by the vendor... I wonder if anyone back then "hacked" their own mainframes?
Er, "their own mainframes"? Surely you know IBM did not sell mainframes for an extreamly long time, just rent them. Sure it was in your machine room. You had to supply power and cooling, and your own experts to run it...but IBM still owned the machine.
I don't know exactly then the practice stopped (I'm not old enough to have gotten in on more then the tail end of the mainframe's era of supremicy...and Amdahl had been around for a while by the time I knew anything!). I'm guessing very late 70s, but more likely the early 80s.
By the way, I think Sun is currently doing this with their high end machines now. Of corse they advertise it. Interesting policy.
Now that the treo has a flash rom, and they're lying about it, what do they expect?
They are lieing about it? Didn't the article cite Handsprings on Knolage Base as saying more or less "sure, we have FLASH now, on some devices, but we don't promise to keep using it".
I think that matches up with the truth.
It may not be a good idea, but that doesn't make it a lie! (it also may be a good idea, if very few people would use the flash, it might be better to make $5 more per unit by switching to MROM and lose some of the people who would have used the flash...of corse if it were my choice I would pay the extra $5 because I would rather patch the OS in ROM then have a RAM patch that a hard reset can take out)
I think you missed my point. Once my browser gets the file it caches it locally. I can retrieve the file from my cache. Can you stop that?
Once you come through the front page to see the item, I'm pretty happy, even if you decide to later look at a cached version, or a print out. I'm less happy if you give the deep link to all your friends and they skip the front page.
Or at least that is the opinions I got from some people that wanted to prevent deep linking a few years ago. For my personal content so long as you link to one of my HTML pages that's fine (my JPGs, not so much). They should all have up links. More over they are all gone since I left UUNET, but that's another story.
Mostly I don't so much care if people deep link to my stuff, or well designed commercial works. It's how the web was designed. Framing another site's stuff though, that is a rip off.
No you won't, not unless you go get a job at apple.
Sure I will. I don't need the source code to the libs to use "dup2(2)" on the mach port file descriptor. It may make it hard to find the fd...except I think there is a call to ask where it is so you can use select on it.
It would require source (or non-trivial amounts of re-implementing) if there was more to it then talking over an fd...but in the past that is really all there was, and it is likely that is all there is now.
At least if you felt like trying to fix XFree86, you could.
I know, I have a patch into X11R4 (which is a distant ancester of XFree86). I wrote the code in twm to use M4 (trivial...except I did at least export some of the X info into M4 macros), which later morphed into tvtwm, and fvwm, and many other window managers.
And you know what? After all the work I and other people put into X11 (and no doubt, it is mostly others), it still sucks 20 times worse the OSX.
It's not that I think free software can't do a good job...it's that I think X11 has done a bad job. A very bad job. And hobbled everything done on top of it. But hey, who am I to say so? I mean all I've done is write a few X apps, and tinker with 20 or so more over the last decade.
I think the best solution is MacOSX's slide-down "dialog sheets" (or whatever they're called).
The generic term is "sheet", they can have property sheets, save sheets, alert sheets, foo sheets.... Or at least that's what they are normally called. Internally they are "NSPanel" classes, or subclasses of them.
KDE, Mozilla, and Gnome can be slow, and some misbehaved applications that don't use mouse grabs properly can make X11 appear to "crash" (it's really working fine, you just need to kill the application--happens under OSX and Windows as well).
I don't know about Windows, but under OSX it is stunningly rare to be unable to switch to another app, or use CMD-Opt-Esc to kill a misbehaving app. I think it may have happend to me once. Maybe. That's fewer times then I have had a kernel panic! (which were all from me doing umount -f as root -- which they seem to have fixed!)
Have you ever had it happen? Are you sure the Dock didn't lock up (that use to happen under 10.0 once in a while, kill it from the terminal and it gets respawned...havn't ever seen it under 10.1)
can quartz natively support network transparency like x11 can?
I doubt it can at the moment, but back whent it was NeXT's DisplayPostScript it definitly could, and did. I use to do the "shooting holes" thing on other people's display at school. Great fun.
Under OSX, if you were to dig deep enough into the frameworks you could probbably get a "MACH port" open to a remote machine's window server (one hopes tunneled over SSH) and there is a good (but not great) chance that it would "just work". Even the old sound APIs were that way. NeXT actually had a way to ask for this though, and Apple doesn't. Of corse so few people did anything at all with it on the NeXT, who blames them for dropping it?
quartz is a step back.
For network transparency, yes. A step forward for anti-aliased text. A step forward in fact for anti-aliased everything. A step forward for using vector based drawing. A step forward for caring about the physical size of rendered objects rather then pixel sizes (rember it's all PostScript inside, even if it is pronounced PDF). Oh, and in gaurenteeing backing store to apps.
That could all be added to X11, but it wouldn't be apps that wanted to use those features would either fail on old X servers, or be six times as complex to write. And adding all that to X11 would take way to long.
Don't beleve me? Well think aobut this, Quartz is what NeXT had in 1990 (1991? 1989?) plus alpha transperency. Why didn't X take the decade and catch up already? Since it didn't, what makes you think Apple should have grabbed X11, and slammed all the wonderful crap the bought from NeXT into it?
(and yes I know about Keith Packards' nice aa extentions to X...but are they done yet? And are they pervasave like they are in Quartz? Oh, and do they solve the other 15 giant gaping voids that X has instead of features?)
If X11 hasn't cought up in a decade, do you think maybe it would be quicker for Apple to be able to make Quartz network transparent then for Apple to help X catch up? Oh....and does Apple's rather expensave "remote desktop" package count?
apple is acting just like all the other proprietary unix vendors did, "look at my nifty proprietary gui!" and if they have any sense they'll just give the fuck up and use x11 like sensible people.
Sure, on the other hand unlike the other Unix vendors so far they seem to be winning. Sure, for reasons other then the rendering technology (it really isn't that much more then NeXT's DPS, or Sun's NeWS!). However the rendering technology is definitly not hurting them.
listen up you little obnoxious x bashing weenies: x11 is a whole lotta baby and an itty-bitty bit of bathwater. don't toss them both; add to the baby and toss the bathwater.
I have written a lot of X apps in my life. Ones that used Xlib directly (xtank for example - no I didn't write all of it, but I was one of the lead maintainers for far too long), ones that used toolkits (Xt and Xaw, Xt and Xmw, Xt and other random crap....GTK--, and others). I know just how big that baby is. If you add more to it, the rest of the bathwater will be forced out of the tub. Of corse you risk the tub busting through the floor too.
I don't hate X. But after writing some small OS X Carbon apps, I really can't keep defending X. I mean Quartz does so much more the X11, and it sure seems faster, and simpler to use. And I expect the network transparency could be fixed. Who knows, maybe I'll poke at that sometime.
sidenote: IBM was a source for G3's, but Motorola was unwilling to license the technology for the G4. Is this still the case?
Not exactly. IBM has G4 class CPUs. In fact the POWER4 is much much faster then any Moto CPU (and more costly), it will run PowerPC code (and POWER code, and PowerAS code, and...). What it won't do is AltiVec, which as far as the Apple world cares is what makes G4 a G4.
Moto was always willing to licence the AltiVec stuff to IBM, but IBM was untilling to pay the fees for quite a while. Then somewhat quietly last year they licenced it. I was hoping htey had put it in the POWER4 and left it disabled, but that clearly didin't happen (and of corse I expect the POWER4 CPU brick itself to cost way more then any existing Mac, so Apple would have to be serious about going into the scientific and DB Unix markets...and IBM would have to decide getting money for the CPUs is worth possabably losing Unix box sales!).
I see a tough road for Apple. If they stick with Motorola, they are subject to Motorola's utter inability to perform.
That is a pretty tough road to hoe. Of corse Moto may have tricks up it's sleave. I think they are using a SOI process for the CPUs, and there is a recent embeded DRAM process for SOI that could let them put 32M of L2 cache on the same chip as the CPU...which could be a serious help for anything that is memory bandwidth bound...as long as OSX's scheduler has processor afinity.
Or Moto could be the same bunch of incompetent boobs who had the 32 bit world on a platter with the 68000, then refused to make RISC chips for Sun, couldn't keep ahead of the x86, and eventually fell totally flat on it's face.
I know which I want...but I know which is more likely. Bummer.
write out a z buffer with the image using -z at the command line.
You can get at the Z buffer, just not from the command line. There is a debugging thing to let you do it, I've used it to try to figure out why some of my scenes didn't come out as I planned. You can also get the alpha chanel rendered directly into PNGs...
and some changes that allow colour gradients to be used for normal gradients.
You can do that now, it looks like 3.5 finaly unifyes all the patterns. Plus you can supply your own functions and us uv-maps!
I've also got a reasonable (but 4 years old!) fractal landscape generator I wrote for POV,
I have a rolercoaster "construction kit" I did about the same time. Maybe we should combine them, since I just put my coaster over water...(and it took me forever to find out that no_shadows also canceled the fake caustics!)
If you know of a way to do this (for every browser, including LWP, wget, etc) let me know.
Trivial, just do it in the web server, not JavaScript or anything the client deals with. The downside is it tends to make it impossable for people to bookmark your pages (normally no referer...some old browesrs sent whatever page they were on before the bookmark).
I wrote this into a CGI script once, but the right place is an apache module, or whatever the equivolent for your web server is. Probbably take less hten a day. In fact since I havn't been keeping up, it might already be in Apache!
I haven't seen the bugtraq posting, but I've read the posting on Macslash, and nowhere does it make the claim that this attack has been proven to work. Instead, the claim is made that because Software Update uses port 80, the attack must be possible. [...] The hole exists _only_ if there is no client-side authentication of what's been downloaded.
The authentication is either non-existent, or very weak. You can
get fake update packages that are really backdoors and the updater will install them if you trick it into taking them. This guy used ARP spoofing which requires you to be on the same physical network. Maybe fairly safely outside the building via 802.11, but still on the same network. Or at least already have cracked another machine on that network.
So yeah, I would say Apple needs to get it's act together and start signing it's stuff, and make the updater support signed packages. If they store the keys in the normal keychain that could even let 3rd parties using Apple's normal installer (assuming you check in the install app, not update!) do "more secure" updates. Of corse the better OSX apps are just "drag into place", and don't use an installer...
The point is that the official stance of the corporation is all that matters in the long run
My point is there is no long run in any comsumer contract with the "we can change it any way we like any time we like" clause, there is no long run. Just think of any contract that says that as having a clause that says "we'll get around to screwing you, just wait".
Don't let a friendly representative fool you into thinking that the corporation is benevolent
Don't let anything fool you into thinking a corporation is benevolent. By law any public corporation can only care about it's shareholder returns. So of corse they aren't benevolent. Well, not to you at any rate.
enforce the agreements you've signed
Or, and here is the real point, the agreement you didn't sign. Since all they need is a little time to change it to whatever they like.
What happens when the friendly guy you talked to gets a better job and the new guy isn't so friendly? Now he has the power to cut you off because you're breaking your contract.
BFD, I'm quite sure there is a "like it or lump it" clause in the contract that lets them change it on 30 days written notice, or the like. So even if NATing was allowed, if the "friendly guy" left, and "mean guy" showed up, it only makes 30 or so days difference.
Of corse it would be nice if doing things like plopping a run of the mill 802.11 access point with the sock config onto your local network were allowed. Even nicer if they would let me pay for a reasonable number of fixed addresses. Not that the DHCP assigned ones seem to ever change. From a practal point of view though, it doesn't matter.
To have the same impact today I think you'd have to have something that made the iMac look ugly and blew away a hefty desktop PC for $300
Um....the Amiga never looked like an indrustrial design masterpiece. From the outside it looked a lot like the other computers of it's era. A more or less rectangler box, muted colors, and some floppy drive slots. Or at least the A1000 and A2000 didn't look special.
What was special last time around was the hardware and the software. This time around the hardware can't be special, so it's the software or bust.
Finally, you can be quite a bit more secure by installing without the BSD tools installed. I dunno if this installs Apache, but it would disable all command line tools. I'm only speculating about that, actually, as I have never done an install like that--as a matter of fact, I rush right out to install the dev tools as soon as possible. But the security minded may want to try it.
I assume you mean stuff like the shell, cp, and mv? You can't really not have those. Some of the GUI stuff could in theory depend on it...and in particular one does.
The apple "package" installer (which is actually pretty good!) will look for several "scripts" during the install, and run them. They are almost always shell scripts, and need the fileutils. Go look at your package reciepts and poke around, you can see some examples. I don't know if any use perl or not, but...
Apple licensed cloning. Around the same time we started seeing demos of the PReP and CHRP boards. These could have run the Mac OS, along with several other operating systems, but to my knowledge no Mac compatible boxes were ever released
Quite a few actually.
the OF could still be a tricky river for an intrepid cloner to navigate
OpenFirmware is oddly enough an open standard. IEEE reference materals and all. ou cna even buy OF implmentations for serveral CPUs from several places...and given that it is mosly a big chunk o FORTH code, it isn't that hard to write, and is normally small. This is not a big problem.
To the extent the cost of equipment is dependent upon volume, this may not be a high enough volume product to make it as a "mass market" board
That's for sure. This Amiga MB is only about $200 cheaper from buying a CRT iMac with roughly the same specs, and that would come with a (funky) case, power supply, speakers, cables, memory, a hard drive, a CD-RW, a small but high quality monitor, and product support.
them. Of course, the G3 and G4 perform comparably per MHz in non-Altivec operations
Well that depends on the G3, and Apple isn't exactly picking G3 machines with the biggest cache and all now, while they are selecting at least some of their G4's with an eye towards speed.
OS X, however, on G3 machines seems rather pokey.
OS X tends to use AltiVec in way more places. Including any of the software rendering, which OS X currently does more of then OS9 because of alpha composing and the like (10.2 has been promised to use the 3D accel pipeline to speed up what you would think of as 2D ops...at least for some video cards). The end result is more OS X stuff is AltiVec stuff.
Finally, I would like to see commodity G4 based boards that could be coaxed to run OS X.
It is easy to get everything but the graphics to work (and maybe audio). For the video you basically need to use the exact same graphics card Apple does. That makes life painful. I think it also breaks the EULA (don't know if it is enforcable) where Apple only gives you the right to run (most of) their software on their hardware. Clearly if they were MS that would be unenforcable, but they aren't. With 5% or less of the market, they can do all the forced product bundling they want.
Slashdot Mirror
Sort of. If they court lets WCOM go into chap 11 the debt is all susspended, along with any pre-chap 11 bills of any kind. To re-emerge from chap-11 they have to show that they are no longer losing money each month, and that they have a "repayment" plan worked out with their creditors. The repayment plan is normally some variant of "screw you, for ever $1 we owed you you get $0.03 and one share of stock...ha ha ha!".
In other words chap 11 basically lets you strip off all the debt (or a huge percentage of it). It also lets you out of any long term contracts (to buy or sell!) that you didn't like, makes it easier to fire people. Oh, and they can "un lease" a building by taking all their stuff out of it and saying "no, were are not paying" to the landlord.
So "too much debt" is no longer WCOMs problem. It is the problem of banks that loaned money, and supplyers that sent goods on purchase orders.
Now I have a theory that the little telco's going bankrupt forced prices down (the little telcos suddonly had "free" infrastructure!) which screwed the medim size ones...and so on. Not just the bankruptcy of corse, it was mostly the drop of the "dot coms", but still I think the bankruptcy laws do have a bit of a chain reaction type thing.
Assuming they are selling a service at a cost that at least covers current costs (ignoring build out a deployment cost!) then either they will keep running it in chap 11 (they can choose to break contracts without peanilty...so they may decide to stop offering a service if they are not making money on it! they can also decide to renogiate any long term contacts they wish, even if they are profitable)....er where was I? Ah, either they will run it in chap 11 and keep running when/if they re-emerge from bankruptcy, or they will sell it off with the infastructure to some company that wants to keep running it.
If the service is not profitable they may try to raise the price, or sell it off, or discontinue it. This is pretty much the golden opertunity for that sort of thing, so if you are getting a "great deal" on something from them, that might not last.
It means if they do let you go they only have to pay out about two weeks of vacation tops as severence. There are also limits on expense reimbursment. On the other hand you paycheck is about as protected as it gets. Unless of corse they lay you off.
It also means they are (I think) less stringent requirments for the layoffs and such. It also means they are prohibited by law from paying off any pre-chap-11 debt, and nobody can deny then continued service on contracts so long as they pay the new bills. Which makes it a lot eaiser to stay in busniess.
Nope, that is just syntax to parse. Complex messy syntax, but still just syntax. Once you parse it, you treat it like a function call.
Templates are more then syntax because use of a template creates a new type on the fly, so your type tracking system has to be able to handle making types when it sees templates.
On the plus side, at least you are not using ObjectaveC which lets you pass messages to objects of a totally opaque type. At least with Java/C++ you always know (er, except with dynamicly loaded objects, which you say are normally rare) that a variable X is of type T, or any descendent of T. With ObjC what you frequently know about variable X is it is "some sort of object or other, maybe that someone wrote code for..." :-)
Oh, I didn't mean to imply overkill so much as if you have a hard problem, sometimes it is easier to solve part of it, and then ask for help, then start off with none solved and ask for help.
I don't know that this is such a case, but it could be!
(personally, I'm less interested in the audit then the function call tree the auditor has to build, and the variable reference tree it should build as well....er, has too if you care about constructor/destructor calls)
That is an interesting approach, but it means your audited libs are not if they expose any virtual functions. Maybe not a big deal. On large projects it means maybe I can't audit all of my code until you audit some of yours, which can be a big deal. It also makes some forward progresses very hard to see.
A long long time ago I wrote a bunch of code analysis tools for APL2 code. For a while I kept getting tempted to re-write them for C. When I learned C++, I just plain gave up. It is so much harder to figure out what function is being called when you just see t->foo()!
I'm quite sure that only the parsing of the Java is easier then C++, not the actual analysis. Er, except maybe for templates. Drat.
Sometimes. Other times the many-heads produce a really complex set of goals, and the project flounders. A single head could produce a realistically small set of goal, write some code to half do the job, and then kick it on out where 50,000 people go "that's stupid, I could do better", and 3 people submit patches to make stuff better.
Pretty much all new code. New data structures. Slightly different view of the problem. Totally different network protocalls. And probbably zero interoperability.
Sounds like a good time to change the name. Even if many of the devlopers are the same.
Java can let you load a class by name and call it. The name can be dynamically constructed. So any function that does this is unauitable. (C++ can normally also do this, but there is no standard way)
In both languages any virtual (or non final) member function might be a call to a never seen before class (if, for example you are a libarary).
I don't think C++ is any harder to analyse in this way, both are quite difficult, and full of big nasty traps. C++ is quite a bit harder to parse though.
FYI, Marlborough's demographic before they adopted the cowboy was pretty much as a "woman's cigerette", i.e. the same demographic as Virginia Slims...
In other words...bad example!
Well, they did actually charge people to go from OS9 to OSX...I think. And to go from 10.0 to 10.1 cost $20 if you coulnd't get to a local store (where they were authorised to give out upgrade CDs for free, and burn new ones if the supply ran out even). Roumor has it the 10.2 upgrade won't be free either. All of those things have likely slowed down adoption a little.
However I think what really has them dragging is not forcing people to upgrade wheather the OS is ready or not. I mean MS is doing whatever they can to force people to pay extra to get their old OSes, and here Apple is handing OS9 out for free on new machines. Even having it pre-installed! Why then even go to the effort of making it work on systems designed after OSX was brought out! Heck Apple even made the iPod work with the old OS9 iTunes. The only new software of any signifigance Apple brought out that is OSX only is iPhoto! (er, iDVD maybe too?)
Clearly Apple isn't doing everything they can to get people to go to OSX. Of corse other then (possably) the charging for money part, I think they are doing everything reasonable to get people to switch.
Er, "their own mainframes"? Surely you know IBM did not sell mainframes for an extreamly long time, just rent them. Sure it was in your machine room. You had to supply power and cooling, and your own experts to run it...but IBM still owned the machine.
I don't know exactly then the practice stopped (I'm not old enough to have gotten in on more then the tail end of the mainframe's era of supremicy...and Amdahl had been around for a while by the time I knew anything!). I'm guessing very late 70s, but more likely the early 80s.
By the way, I think Sun is currently doing this with their high end machines now. Of corse they advertise it. Interesting policy.
They are lieing about it? Didn't the article cite Handsprings on Knolage Base as saying more or less "sure, we have FLASH now, on some devices, but we don't promise to keep using it".
I think that matches up with the truth.
It may not be a good idea, but that doesn't make it a lie! (it also may be a good idea, if very few people would use the flash, it might be better to make $5 more per unit by switching to MROM and lose some of the people who would have used the flash...of corse if it were my choice I would pay the extra $5 because I would rather patch the OS in ROM then have a RAM patch that a hard reset can take out)
Once you come through the front page to see the item, I'm pretty happy, even if you decide to later look at a cached version, or a print out. I'm less happy if you give the deep link to all your friends and they skip the front page.
Or at least that is the opinions I got from some people that wanted to prevent deep linking a few years ago. For my personal content so long as you link to one of my HTML pages that's fine (my JPGs, not so much). They should all have up links. More over they are all gone since I left UUNET, but that's another story.
Mostly I don't so much care if people deep link to my stuff, or well designed commercial works. It's how the web was designed. Framing another site's stuff though, that is a rip off.
Sure I will. I don't need the source code to the libs to use "dup2(2)" on the mach port file descriptor. It may make it hard to find the fd...except I think there is a call to ask where it is so you can use select on it.
It would require source (or non-trivial amounts of re-implementing) if there was more to it then talking over an fd...but in the past that is really all there was, and it is likely that is all there is now.
I know, I have a patch into X11R4 (which is a distant ancester of XFree86). I wrote the code in twm to use M4 (trivial...except I did at least export some of the X info into M4 macros), which later morphed into tvtwm, and fvwm, and many other window managers.
And you know what? After all the work I and other people put into X11 (and no doubt, it is mostly others), it still sucks 20 times worse the OSX.
It's not that I think free software can't do a good job...it's that I think X11 has done a bad job. A very bad job. And hobbled everything done on top of it. But hey, who am I to say so? I mean all I've done is write a few X apps, and tinker with 20 or so more over the last decade.
The generic term is "sheet", they can have property sheets, save sheets, alert sheets, foo sheets.... Or at least that's what they are normally called. Internally they are "NSPanel" classes, or subclasses of them.
I don't know about Windows, but under OSX it is stunningly rare to be unable to switch to another app, or use CMD-Opt-Esc to kill a misbehaving app. I think it may have happend to me once. Maybe. That's fewer times then I have had a kernel panic! (which were all from me doing umount -f as root -- which they seem to have fixed!)
Have you ever had it happen? Are you sure the Dock didn't lock up (that use to happen under 10.0 once in a while, kill it from the terminal and it gets respawned...havn't ever seen it under 10.1)
I doubt it can at the moment, but back whent it was NeXT's DisplayPostScript it definitly could, and did. I use to do the "shooting holes" thing on other people's display at school. Great fun.
Under OSX, if you were to dig deep enough into the frameworks you could probbably get a "MACH port" open to a remote machine's window server (one hopes tunneled over SSH) and there is a good (but not great) chance that it would "just work". Even the old sound APIs were that way. NeXT actually had a way to ask for this though, and Apple doesn't. Of corse so few people did anything at all with it on the NeXT, who blames them for dropping it?
For network transparency, yes. A step forward for anti-aliased text. A step forward in fact for anti-aliased everything. A step forward for using vector based drawing. A step forward for caring about the physical size of rendered objects rather then pixel sizes (rember it's all PostScript inside, even if it is pronounced PDF). Oh, and in gaurenteeing backing store to apps.
That could all be added to X11, but it wouldn't be apps that wanted to use those features would either fail on old X servers, or be six times as complex to write. And adding all that to X11 would take way to long.
Don't beleve me? Well think aobut this, Quartz is what NeXT had in 1990 (1991? 1989?) plus alpha transperency. Why didn't X take the decade and catch up already? Since it didn't, what makes you think Apple should have grabbed X11, and slammed all the wonderful crap the bought from NeXT into it?
(and yes I know about Keith Packards' nice aa extentions to X...but are they done yet? And are they pervasave like they are in Quartz? Oh, and do they solve the other 15 giant gaping voids that X has instead of features?)
If X11 hasn't cought up in a decade, do you think maybe it would be quicker for Apple to be able to make Quartz network transparent then for Apple to help X catch up? Oh....and does Apple's rather expensave "remote desktop" package count?
Sure, on the other hand unlike the other Unix vendors so far they seem to be winning. Sure, for reasons other then the rendering technology (it really isn't that much more then NeXT's DPS, or Sun's NeWS!). However the rendering technology is definitly not hurting them.
I have written a lot of X apps in my life. Ones that used Xlib directly (xtank for example - no I didn't write all of it, but I was one of the lead maintainers for far too long), ones that used toolkits (Xt and Xaw, Xt and Xmw, Xt and other random crap....GTK--, and others). I know just how big that baby is. If you add more to it, the rest of the bathwater will be forced out of the tub. Of corse you risk the tub busting through the floor too.
I don't hate X. But after writing some small OS X Carbon apps, I really can't keep defending X. I mean Quartz does so much more the X11, and it sure seems faster, and simpler to use. And I expect the network transparency could be fixed. Who knows, maybe I'll poke at that sometime.
Not exactly. IBM has G4 class CPUs. In fact the POWER4 is much much faster then any Moto CPU (and more costly), it will run PowerPC code (and POWER code, and PowerAS code, and...). What it won't do is AltiVec, which as far as the Apple world cares is what makes G4 a G4.
Moto was always willing to licence the AltiVec stuff to IBM, but IBM was untilling to pay the fees for quite a while. Then somewhat quietly last year they licenced it. I was hoping htey had put it in the POWER4 and left it disabled, but that clearly didin't happen (and of corse I expect the POWER4 CPU brick itself to cost way more then any existing Mac, so Apple would have to be serious about going into the scientific and DB Unix markets...and IBM would have to decide getting money for the CPUs is worth possabably losing Unix box sales!).
That is a pretty tough road to hoe. Of corse Moto may have tricks up it's sleave. I think they are using a SOI process for the CPUs, and there is a recent embeded DRAM process for SOI that could let them put 32M of L2 cache on the same chip as the CPU...which could be a serious help for anything that is memory bandwidth bound...as long as OSX's scheduler has processor afinity.
Or Moto could be the same bunch of incompetent boobs who had the 32 bit world on a platter with the 68000, then refused to make RISC chips for Sun, couldn't keep ahead of the x86, and eventually fell totally flat on it's face.
I know which I want...but I know which is more likely. Bummer.
You can get at the Z buffer, just not from the command line. There is a debugging thing to let you do it, I've used it to try to figure out why some of my scenes didn't come out as I planned. You can also get the alpha chanel rendered directly into PNGs...
You can do that now, it looks like 3.5 finaly unifyes all the patterns. Plus you can supply your own functions and us uv-maps!
I have a rolercoaster "construction kit" I did about the same time. Maybe we should combine them, since I just put my coaster over water...(and it took me forever to find out that no_shadows also canceled the fake caustics!)
Trivial, just do it in the web server, not JavaScript or anything the client deals with. The downside is it tends to make it impossable for people to bookmark your pages (normally no referer...some old browesrs sent whatever page they were on before the bookmark).
I wrote this into a CGI script once, but the right place is an apache module, or whatever the equivolent for your web server is. Probbably take less hten a day. In fact since I havn't been keeping up, it might already be in Apache!
The authentication is either non-existent, or very weak. You can get fake update packages that are really backdoors and the updater will install them if you trick it into taking them. This guy used ARP spoofing which requires you to be on the same physical network. Maybe fairly safely outside the building via 802.11, but still on the same network. Or at least already have cracked another machine on that network.
So yeah, I would say Apple needs to get it's act together and start signing it's stuff, and make the updater support signed packages. If they store the keys in the normal keychain that could even let 3rd parties using Apple's normal installer (assuming you check in the install app, not update!) do "more secure" updates. Of corse the better OSX apps are just "drag into place", and don't use an installer...
My point is there is no long run in any comsumer contract with the "we can change it any way we like any time we like" clause, there is no long run. Just think of any contract that says that as having a clause that says "we'll get around to screwing you, just wait".
Don't let anything fool you into thinking a corporation is benevolent. By law any public corporation can only care about it's shareholder returns. So of corse they aren't benevolent. Well, not to you at any rate.
Or, and here is the real point, the agreement you didn't sign. Since all they need is a little time to change it to whatever they like.
BFD, I'm quite sure there is a "like it or lump it" clause in the contract that lets them change it on 30 days written notice, or the like. So even if NATing was allowed, if the "friendly guy" left, and "mean guy" showed up, it only makes 30 or so days difference.
Of corse it would be nice if doing things like plopping a run of the mill 802.11 access point with the sock config onto your local network were allowed. Even nicer if they would let me pay for a reasonable number of fixed addresses. Not that the DHCP assigned ones seem to ever change. From a practal point of view though, it doesn't matter.
Um....the Amiga never looked like an indrustrial design masterpiece. From the outside it looked a lot like the other computers of it's era. A more or less rectangler box, muted colors, and some floppy drive slots. Or at least the A1000 and A2000 didn't look special.
What was special last time around was the hardware and the software. This time around the hardware can't be special, so it's the software or bust.
I assume you mean stuff like the shell, cp, and mv? You can't really not have those. Some of the GUI stuff could in theory depend on it...and in particular one does.
The apple "package" installer (which is actually pretty good!) will look for several "scripts" during the install, and run them. They are almost always shell scripts, and need the fileutils. Go look at your package reciepts and poke around, you can see some examples. I don't know if any use perl or not, but...
The normal boot process may well also need them.
Quite a few actually.
OpenFirmware is oddly enough an open standard. IEEE reference materals and all. ou cna even buy OF implmentations for serveral CPUs from several places...and given that it is mosly a big chunk o FORTH code, it isn't that hard to write, and is normally small. This is not a big problem.
That's for sure. This Amiga MB is only about $200 cheaper from buying a CRT iMac with roughly the same specs, and that would come with a (funky) case, power supply, speakers, cables, memory, a hard drive, a CD-RW, a small but high quality monitor, and product support.
Well that depends on the G3, and Apple isn't exactly picking G3 machines with the biggest cache and all now, while they are selecting at least some of their G4's with an eye towards speed.
OS X tends to use AltiVec in way more places. Including any of the software rendering, which OS X currently does more of then OS9 because of alpha composing and the like (10.2 has been promised to use the 3D accel pipeline to speed up what you would think of as 2D ops...at least for some video cards). The end result is more OS X stuff is AltiVec stuff.
It is easy to get everything but the graphics to work (and maybe audio). For the video you basically need to use the exact same graphics card Apple does. That makes life painful. I think it also breaks the EULA (don't know if it is enforcable) where Apple only gives you the right to run (most of) their software on their hardware. Clearly if they were MS that would be unenforcable, but they aren't. With 5% or less of the market, they can do all the forced product bundling they want.