There's been a lot of talk to the effect that Apple is not likely to abandon the catchy-sounding "OS X" name. ("O S X I" doesn't sound as cool as "O S X"....) So will they call it "OS X Eleven" or "OS X Two point Oh" or what? Who knows?! As much as the OS might deserve a full new version number, the marketing aspect of it definitely pulls in the direction of keeping "OS X" as long as possible.
I'd think this would be an ideal branding opportunity;
OS X
i
...to promote the i-mac-product-of-the-month branding. It would sound 'cool' and sophisticated similar to made up car names; 'oh-ess ecks-eye'.
Also, our management goes by the phone system time for scheduling, while the IT department forces a sync with its servers on login, which are always at least 5 minutes behind the phone system. This messes with our calendars and meeting notifications. Maybe we shouldn't depend on Outlook to tell us when our next meeting is, but we all do here. If I'm not allowed to set my PC clock to the correct time, that slows down daily operations to an extent, though I admit it's mainly an inconvenience.
Which clock is synced to the official time for your area?
The phone system
The network
Both should be the same, within at most a few seconds, not 1, 2,... or 5 minutes off. Both have mechanisms to sync to 'atomic time' automatically. If your phone and network admins can't be bothered to do something about that...the problem you have is social not technical.
This is not quantum science folks. If the checksum (typically MD5) does not match what you expect, delete the file. If no checksum exists, ask for one!
I hate math. When I was young, any joy on the subject was tourtured out of me by drill sessions at school and an impatient parent who kept jumping ahead of me with the answer.
That said, "Rapid Math Tricks and Tips: Thirty Days to Number Power" by Edward H. Julius is great. It is a little cheesy, but very practical. It allows you to do much of the same calculations that a 'child prodgy' can even if you're old.
It does not help with number theory, though it can help give you a much better feel for numbers.
It's just become another standard - not the de facto one. De facto implies that it is, in fact, the standard, as opposed to, say, de jure, which is a legal standard (cf. "laws more honoured in the breach").
*grumble*
A nit: Standards are based on specifications. Implementations of specs are called 'Standard' by convention but are implementations, not standards. The implementation can shape an existing standard or even create a de-facto standard.
A popular de-facto standard is Microsoft's.doc format that is documented incompletely or impractically since even Word does not implement the format the same way the documents describe or even consistantly in different implementations of Word.
Other programs that also implement the de-facto standard.doc format are variably sucessful in creating and/or reading.doc files created using Microsoft's implementation.
With the increase in viruses, the chance of flase positives rises substantially.
With the false positives, and the delays in identifying new viri, the virus writers have a gap in time and a gap in confidence that the detector is both timely and accurate.
For example McAfee Virus Scan reports a false positve for a development file SetCVSShellCommands.exe from NSIS (the installation program used by WinAmp and other programs). McAfee has been unresponsive on adapting Virus Scan so that it does not trigger a false positve. All they are interested in is actual viruses or -- in this misidentification -- spy ware.
I doubt that SetCVSShellCommands.exe -- basically NSIS itself with a custom script -- will be handled properly anytime soon.
This is not a new problem, though. One program I worked on about 10+ years ago was identified as a virus because it had a jump to a BIOS location and the string 123 in it (software reboot). At the time, the false positive was handled promptly and was not an issue when the virus detection software was updated.
All your talk about "customer management" just makes you sound like some slick-talking dotcom sales bullshitter and not someone who deliveres solutions for a living. Get a taste of reality, please.
Brave words from an AC. Maybe I'm not the one who needs to learn about this subject??
Your example shows what not to do; ignore the customer. They want protection, you talk about cheap. Maybe they are right or you didn't talk to them properly? I can't tell...no specs, no requirements!
I hate to disagree with that, but we recently had a project and had the choice of which SQL to use. Customer pushed back and simply said MS.
I'm sorry...I do understand what you're talking about and in some ways I agree though I'm angry and need to vent. In short;
Your main problem is customer management and requirements...not the tool chosen.
The customer sets expectations (the what) and your job is implementation (the how).
When the customer specifies the tools and the background methods they are saying they know better and could do your job. As if they actually know wtf you do (*sprinkle bitterness*).
A note on 'customer management'. Also known as "managing customer expectations", managing the customer is something that has to be done if you want a sucessful project most of the time. It gets a bad name because some see it as a method to abuse or ignore customer complaints. That is the wrong way to do it. Both the customer and the implementor -- even if you are an in-house developer or tech -- have roles.
The customer knows what they need (in general) but does not have the expert knowledge to do the work.
You should know how to do the work and to ask the questions needed to implement a system to support those needs.
Do you tell a plumber the type of pipe he has to use, or do you talk with the plumber and understand that he knows wtf he's talking about and then (if he does) let him do his job?
What is it with people thinking they are in the software business when they are really in a business that just happens to use software? Everything will be computerized, just like everything uses electricity...though not everyone is an electrician!
They did for a huge chunk of the earlier drivers, yes. IIRC, Windows 95's drivers were initially all written by Microsoft - or at least a huge chunk of them were.
Not to be too harsh, though, how many months did that last?
First, its it's an intestinal irritant. Its sold in gelatine capusules and if you just swallow one a day, you'll be sorry after a while. I recommend opening the capsules and disolving the choline in something buffered, like milk.
Depends on the form/packaging. Solid and powder forms are available also. These don't cause me any problems, though I have iron pipes and can eat anything short of poison or tin cans.
It cuts the kernel stack down to a single 4k page. There are performance advantages to doing this.
The main caveat is that drivers (and other kernel code, really) have to be very careful about their stack usage, or they run out of stack.
Consequently, badly written binary drivers with fat glue layers are right out...
Yep, I could see that one comming. Having worked at a company that used to do ye-olde-386-DOS-Memory-managers (386MAX), I realize the performance benifits of these types of tweaks.
I'd expect that Nvidia will have to spend quite a bit of effort to adapt to this one even if it only is used in Fedora (and that I doubt).
An OSS X driver would be better. I'll have to choose carefully when getting a laptop.
Nice shot at trolling, but better luck next time...
Was that really necessary? As the rest of your comments are more reasonable, I'll reply to those.
Like, why was Microsoft talking to anyone about funding for TSG? How can a monopoly justify interfering in any way attacks on possible competitors?
Law and morality aren't the same thing.
Like, what difference does it make if the executive was not BG or SB? Those are not the only executives at MSFT, and they are not the only executives capable of committing MSFT to some kind of deal with BayStar.
It doesn't to me; these were statements from Baystar.
Like, according to a link posted somewhere above (this one) Microsoft was one of the ten top investors in BayStar, so any investment by BayStar MAY have a Microsoft thumbprint on it.
Gotcha. Thanks for the fact (first one so far from anyone on this thread).
Like "A Microsoft spokesman says that the company has no "direct or indirect" financial relations with BayStar" despite the pdf file linked to above found on BayStar's web site.
That would follow. I'd like to see if they back pedal on that one or just leave the lie(s) out there.
A nice unbiased article about how Linux is superior...from a Linux magazine. Perhaps we'll be posting the article from Windows Insider about how Windows is better? No? Didn't think so.
Note that he generally talks about Unix and when being specific most of the time Solaris is mentioned. When he talks about *BSD and Linux it is along the lines of 'oh, and other Unix-like systems such as...'.
Now, anyone who has read the Linux kernel list or at a minimum Kernel Traffic will recognize that what he says does indeed apply to Linux as well. Without that knowledge, though, it looks like the reasons why Solaris is better and Linux just happens to be like Solaris. (As someone who has admined solaris boxen and owns a few ancient ones -- I'll take Solaris though I want Linux if at all possible.)
That could very well be true. But it doesn't mean an investment could not have been made by a shell company of Microsoft.
Microsoft -> shell company -> Baystar
That shell company could be another company like Paul Allen's Vulcan Capital.
Agreed (and entirely expected from MS's history). Is there any way to check this?
That said, though, I asked some friends if this were legal and ethical. One that I trust to be level headed, well informed, and wise pointed out;
Legal: Yes.
Ethical: Nope.
Another -- a blind zealot for everything MS; even other staunch MS advocates tend to give him odd looks -- mentioned that not only was it appropriate for MS to throw money at knocking down Linux (and IBM and Novel and...) any way they could Microsoft is responsible to do so as defensive tactic as are any other companies threatend with competition.
His recovery from calling Linux competition was quick, though, as he pointed out that Linux isn't really competition just an annoyance and a distraction. {BSE ~grin~}
I'm not surprised that people are having problems with 2.6.x, though I think the 99.9% you mention isn't reasonable or realistic either! (The Windows updates you mention are for complete shipping versions that have been out for a couple years; the updates shouldn't cause unreasonable problems.
That said...
...the kernel doesn't need to be updated. It is not an application; it is support.
In a production environment do not play with existing servers that use 2.4.x. Use a test machine (or your own system) with 2.6.x. When the distribution you use is upgraded, test *that* and upgrade if it fits your needs. Otherwise, don't!
That said, I've only had problems with Fedora Core 2 test 1 and Nvidia's binary drivers. (Not that that surprises me at all.) Everything else is damn stable, and one annoying issue with USB mass storage crashes went away entirely (supposedly also fixed in a later version of 2.4.x).
I've had very little trouble with 2.6.x, either the custom one I whipped up a few weeks ago or the binary version distributed with Fedora Core 2 (test 1). The updates seem to work well also.
The only nit: The CONFIG_4K_STACK addition that is in FC Core2 test kernels but not the official kernel.org sources is incompatable with Nvidia's binary drivers. Getting the config settings for the current kernel from/boot/ and rolling your own seems to be the current workaround.
Anyone know what CONFIG_4K_STACK does? I can guess from the name, though that's not knowledge.
Slashdot Mirror
I'd think this would be an ideal branding opportunity;
...to promote the i-mac-product-of-the-month branding. It would sound 'cool' and sophisticated similar to made up car names; 'oh-ess ecks-eye'.
Neither do I here. The network is right, but the phone system has 'official project time' (about 6 minutes fast). Drives me nuts...
Well, there goes my one example of perfect code.'
Which clock is synced to the official time for your area?
The phone system
The network
Both should be the same, within at most a few seconds, not 1, 2, ... or 5 minutes off. Both have mechanisms to sync to 'atomic time' automatically. If your phone and network admins can't be bothered to do something about that...the problem you have is social not technical.
FK! The new rule at /. is to undo moderation even if you post as AC!
This is not quantum science folks. If the checksum (typically MD5) does not match what you expect, delete the file. If no checksum exists, ask for one!
That said, "Rapid Math Tricks and Tips: Thirty Days to Number Power" by Edward H. Julius is great. It is a little cheesy, but very practical. It allows you to do much of the same calculations that a 'child prodgy' can even if you're old.
It does not help with number theory, though it can help give you a much better feel for numbers.
*grumble*
A nit: Standards are based on specifications. Implementations of specs are called 'Standard' by convention but are implementations, not standards. The implementation can shape an existing standard or even create a de-facto standard.
A popular de-facto standard is Microsoft's .doc format that is documented incompletely or impractically since even Word does not implement the format the same way the documents describe or even consistantly in different implementations of Word.
Other programs that also implement the de-facto standard .doc format are variably sucessful in creating and/or reading .doc files created using Microsoft's implementation.
With the false positives, and the delays in identifying new viri, the virus writers have a gap in time and a gap in confidence that the detector is both timely and accurate.
For example McAfee Virus Scan reports a false positve for a development file SetCVSShellCommands.exe from NSIS (the installation program used by WinAmp and other programs). McAfee has been unresponsive on adapting Virus Scan so that it does not trigger a false positve. All they are interested in is actual viruses or -- in this misidentification -- spy ware.
I doubt that SetCVSShellCommands.exe -- basically NSIS itself with a custom script -- will be handled properly anytime soon.
This is not a new problem, though. One program I worked on about 10+ years ago was identified as a virus because it had a jump to a BIOS location and the string 123 in it (software reboot). At the time, the false positive was handled promptly and was not an issue when the virus detection software was updated.
Some Windows drivers do allow multi-user logins, though they are very narrowly taylored and the last one I noticed was for Windows 98/ME only.
With X, you can have a few dozen exported X terms from the same machine and run them locally (multiple cards) or remotely (network).
Brave words from an AC. Maybe I'm not the one who needs to learn about this subject??
Your example shows what not to do; ignore the customer. They want protection, you talk about cheap. Maybe they are right or you didn't talk to them properly? I can't tell...no specs, no requirements!
I'm sorry...I do understand what you're talking about and in some ways I agree though I'm angry and need to vent. In short;
Your main problem is customer management and requirements...not the tool chosen.
The customer sets expectations (the what) and your job is implementation (the how).
When the customer specifies the tools and the background methods they are saying they know better and could do your job. As if they actually know wtf you do (*sprinkle bitterness*).
A note on 'customer management'. Also known as "managing customer expectations", managing the customer is something that has to be done if you want a sucessful project most of the time. It gets a bad name because some see it as a method to abuse or ignore customer complaints. That is the wrong way to do it. Both the customer and the implementor -- even if you are an in-house developer or tech -- have roles.
The customer knows what they need (in general) but does not have the expert knowledge to do the work.
You should know how to do the work and to ask the questions needed to implement a system to support those needs.
Do you tell a plumber the type of pipe he has to use, or do you talk with the plumber and understand that he knows wtf he's talking about and then (if he does) let him do his job?
What is it with people thinking they are in the software business when they are really in a business that just happens to use software? Everything will be computerized, just like everything uses electricity...though not everyone is an electrician!
Not to be too harsh, though, how many months did that last?
Depends on the form/packaging. Solid and powder forms are available also. These don't cause me any problems, though I have iron pipes and can eat anything short of poison or tin cans.
The main caveat is that drivers (and other kernel code, really) have to be very careful about their stack usage, or they run out of stack.
Consequently, badly written binary drivers with fat glue layers are right out...
Yep, I could see that one comming. Having worked at a company that used to do ye-olde-386-DOS-Memory-managers (386MAX), I realize the performance benifits of these types of tweaks.
I'd expect that Nvidia will have to spend quite a bit of effort to adapt to this one even if it only is used in Fedora (and that I doubt).
An OSS X driver would be better. I'll have to choose carefully when getting a laptop.
boxen See the second definition.
Was that really necessary? As the rest of your comments are more reasonable, I'll reply to those.
Law and morality aren't the same thing.
It doesn't to me; these were statements from Baystar.
Gotcha. Thanks for the fact (first one so far from anyone on this thread).
That would follow. I'd like to see if they back pedal on that one or just leave the lie(s) out there.
Note that he generally talks about Unix and when being specific most of the time Solaris is mentioned. When he talks about *BSD and Linux it is along the lines of 'oh, and other Unix-like systems such as ...'.
Now, anyone who has read the Linux kernel list or at a minimum Kernel Traffic will recognize that what he says does indeed apply to Linux as well. Without that knowledge, though, it looks like the reasons why Solaris is better and Linux just happens to be like Solaris. (As someone who has admined solaris boxen and owns a few ancient ones -- I'll take Solaris though I want Linux if at all possible.)
Microsoft -> shell company -> Baystar That shell company could be another company like Paul Allen's Vulcan Capital.
Agreed (and entirely expected from MS's history). Is there any way to check this?
That said, though, I asked some friends if this were legal and ethical. One that I trust to be level headed, well informed, and wise pointed out;
Legal: Yes.
Ethical: Nope.
Another -- a blind zealot for everything MS; even other staunch MS advocates tend to give him odd looks -- mentioned that not only was it appropriate for MS to throw money at knocking down Linux (and IBM and Novel and ...) any way they could Microsoft is responsible to do so as defensive tactic as are any other companies threatend with competition.
His recovery from calling Linux competition was quick, though, as he pointed out that Linux isn't really competition just an annoyance and a distraction. {BSE ~grin~}
How wonderfully naive you are!
While I am wonderful and also naive, I suppose when your knee jerked you missed the next sentence;
OK, am I missing something or is this a non-story?
(Yes, I think that something slimy is actually happening...though evidence of it would be good!)
I've already looked for the quote in the reasonable places, and didn't find it. Anyone have a reference?
That said...
In a production environment do not play with existing servers that use 2.4.x. Use a test machine (or your own system) with 2.6.x. When the distribution you use is upgraded, test *that* and upgrade if it fits your needs. Otherwise, don't!
That said, I've only had problems with Fedora Core 2 test 1 and Nvidia's binary drivers. (Not that that surprises me at all.) Everything else is damn stable, and one annoying issue with USB mass storage crashes went away entirely (supposedly also fixed in a later version of 2.4.x).
The only nit: The CONFIG_4K_STACK addition that is in FC Core2 test kernels but not the official kernel.org sources is incompatable with Nvidia's binary drivers. Getting the config settings for the current kernel from /boot/ and rolling your own seems to be the current workaround.
Anyone know what CONFIG_4K_STACK does? I can guess from the name, though that's not knowledge.