While I agree that the state of C++ compilers could be a lot better (more power to EDG!), the Mozilla C++ Portability Guide is seriously outdated. It is by now 4 years old, and it shows. The claims about the HP compiler, for instance, apply to the previous compiler. No, not just the previous version, the previous compiler, for which HP has discontinued all support some time ago already. Actually, even back in April 98, HP's CC compiler was already being phased out and replaced by the much much better aCC one. Actually, we made the switch some time (I don't recall exactly) before the start of my current project in September of 1997.
Whilst it is true that I think Microsoft is Vile, if it saves all of the threatened clubs my opinion of them will increase dramatically.
You don't really think MicroSoft (or any other big company for that matter) would buy the TV station to save the local football clubs, I hope... Big companies like that don't care at all about such things. As a matter of fact, neither do the vast majority of smaller ones.
To let your opinion of a company be influenced by this kind of "accidental" benefit is a recipe for serious disillusionement.
Looking at the site of my employer, browser statistics depend a lot on the content of the site, but apparently not as much as I thought.
We're a micro-electronics research centre (covering everything from MPEG-4 standard definition, over ECAD, chip packaging, down to our very own cleanroom). This means that we should have more Unix based visitors than the average me-and-my-cat home page. Our stats confirm his, but still MSIE takes between 80 and 90% (the figures vary on a daily basis), which is way more than I was expecting.
Fortunately, as of this week, we finally support all gecko based browsers. Hopefully, they will now start to rise above the disappointing 0.9% they got so far.
The problem with these fake browser strings, is that it helps to keep the Mozilla statistics down, thus giving the PHBs a reason why Mozilla support isn't a priority. That's exactly what I'm facing right now. Our web site doesn't work with anything except IE5+ and NS4 (due to crappy HTML). It took me a lot of lobbying to get people to work on fixing it at least partly, and even now I have to do part of it myself (it's not my job) because if they get stuck, chances are that they'll use the low percentage of Mozilla and Opera visitors to abort the action.
People in the know will know that fake entries exist, but those are so few and far between that it doesn't really matter. Besides, I doubt many admins look at individual entries. They most likely just look at automatically generated summaries. So I'd prefer all those who play games with this, to stop doing it except in exceptional circumstances.
If a site doesn't even let you in, it's not worth your visit. Our site is badly broken too at the moment, but it at least allows people with non-mainstream browsers to enter end get st part of the info.
Yes, the building plans exist and need to be reviewed. However
The architect didn't know the server room layout, so that's not on the "available" plans.
If he designed 5 or 10 identical blocks next to each other, he sure didn't know which one would end up holding A.
Neither did he know all the details of any security mechanism that isn't part of the core structure of the buildings.
Who says that these plans (still) are all that available in the first place? I would be surprised if I could just go to town hall and get the complete plans to one of the local bank offices, even if they have them. ("No, no, Mrs. Clarck Offis, I'm not planning to rob them. I just want to see if it could be done. Yes, I'll warn them if needed.":-)
You can bet that specific regulations exists to deal with the process of constructing security sensitive buildings. In fact, being a reserve naval officer for my country, I can assure you that this is the case over here and I would be very very surprised to find the US more relax in this area.
Because opening up the plans removes the obscurity aspect without adding anything new to replace it. With open source, the white hats can actually fix the problems they see. With open building plans there's nothing Joe Ramdom Hacker can do to actually help improve security. So why would he care?
Also, building security is much easier and (most of all) after thousands of years of practice also much much better understood than software security.
All in all, in the building context, the risk that there are unknown flaws is much lower and the risk that only the black hats (who themselves happily operate under maximum obscurity) would benefit from an "open plan" is much larger.
Ever since so many years ago I had an eye infection and couldn't look at a lot of white light for hours on end, it is also exactly my default background color. At last I know why I've never gone back once the infection was cured: I have by pure coincidence stumbled onto nirwana.:-)
I tried to look at it with Mozilla 0.9.8 on Linux. The thing entered an endless loop, redisplaying the page over and over again. When I hit the back button, mozilla even crashed. So much for a working example site...
And before I forget: I'm actually using an HP-UX desktop (not for long anymore, but still). =Last time I looked, there still was no flash plugin for HP-UX. If Macromedia wants people to sort of drop HTML in favour of Flash, they better start by making their stuff available on ALL platforms, no matter how obscure.
I have a big pile of dead tree books about computing (in the broadest sense of the word) and do tend to look for more interesting ones whenever I have time (which doesn't happen a lot lately, however).
The one key characteristic of most of these books, is that they are (more or less) timeless. The kind of stuff that you can buy today and still use as a reference 5 or 10 years from now. In other words, I want books about concepts, not implementations.
I don't mind if any example code is written in a language that is slightly less hip when I'll be reading it than it used to be when it was written. Hell, I'll gladly buy an earlier edition that is 5 years old and uses what by now already are "all the wrong example systems" if I think the book has intrinsic value on the conceptual front and the new edition basically just translates the examples to something more modern.
What I'll almost never buy, is the kind of computer books that just teach you how to use today's (or rather yesterday's) version of program X or trendy API Y. Those I consider utter garbage. By the time you carried them home, they're hopelessly outdated.
I must admit to buying a few language books in the past, but even then nothing of the "Using Visual Sh.t version 3.1415" or the "Become a Java Pro in 30 Days" kind. To teach myself C++, for instance, I first read a copy of Stroustrup's second edition, which I got from a library, once. Linearly, that is. Then I started out for real by buying the ARM and the Coplien book. I bought a few more advanced books about C++ later, but basically those were not essential for me. By now, if I'd need to update my C++ knowledge to cover the STL more in depth, I'd "simply" dig up my aging copy of the draft standard.
Indeed! xhpcalc is the best calculator program I have ever seen. By a long margin.
HP would make themselves very popular by releasing the source code for that thing, considering that they dropped it as long ago as the release of HP-UX 10.0 and that even so people still preserve it when installing new machines or OS versions.
Unfortunately, I suspect that xhpcalc literally emulates the calculators in question using their original code ROM images. That probably means it's very unlikely to ever see the light of day again, even if they can still find a backup tape containing the source.
Yes, it exists, we have it installed. But in a multi-platform environment, you want to support one browser and one browser only. Netscape is the only option in that case. Besides, you need HP-UX 11 to get a non-beta version of IE for HP-UX.
But all that was not my point. I can also use an NT PC if really needed. My point was that HP is giving a rather special message in doing things this way. A message that leaves me unwilling to believe some of the other stuff they claim.
It would be nice if this commitment would not be undermined by other things they do. For instance, go to this HP-UX server configuration page and try to configure one. You don't need to log in, you can do it as a guest user. Or at least, you can do it if your browser is deemed worthy. Right now, I'm using Netscape on an HP-UX box, but even so it refuses to my business into consideration.
I sure am hoping that this is not the way their Linux commitment will be implemented...
Europeans,... cannot understand what real democracy is about.
Says an inhabitant of a country that doesn't even have something as basic a constitution, in which the head of state is not elected, and in which parties that get 40% of the votes of the people routinely end up with 60% of the seats in parliament.
Re:One simple reason why it won't work:
on
The Euro
·
· Score: 2, Insightful
I would suggest that you try thinking though the context of your statements before engaging in ignorant-Europe FUD.
Many more people understand and even speak multiple languages here in Europe than the average American can even start to imagine. And the percentage of people who don't do so is falling all the time.
Besides, it's not because people in Detroit and Dallas all speak American english that they move around just as easily as you imply. There are many more reasons than language to stick around in the same area.
I really wonder where you got that bit of FUD about the alloys. Everything is normed to be exactly the same, with the single exception of the back side of the coins. It simply has to be, otherwise vending machines would not work at all.
The only issues are whether all producers are capable of sticking to the allowed tolerances and whether those who programmed the vending machines didn't stick to closely to the official norm, ignoring (part of) the tolerances. The first issue should be a non-issue, I think. After all, it alos applied to the previous national coins. The second one can be corrected if needed.
That's because today's mechanic can easily fix yesteryear's car without requiring tons of additional training (finding spare parts might be a different thing though).
Another problem with this car analogy, is that old cars break over time as they get used. So in the end you need to replace them, thus allowing Honda & Co. to live. Old software doesn't break (remember, we're talking binaries here, not source code that is being patched up beyond recognition). It might be broken from day one, but if it was good enough for people to use anyway, it remains so until people's needs change. Hence, the software company needs an other way to keep the money flowing if it itself wants to stay alive.
This is why Microsoft is changing the licensing rules as of XP. Even they can't keep on bloating Word & Co. indefinitely. At some point nobody buys the new version anymore. Unless some licensing mechanism or a dependency on an external component forces them.
Furthermore, any person IN THE COMPANY who sends an attachment to another person in the
company that's rejected by the mail server because the recipient hadn't filled in the form has his or
her email account locked for 24 hours to stop the virus from spreading.
I can't believe I'm reading this. And even less that someone moderated this up as being interesting. Sarcastic would have been more appropriate. Except that it likely wasn't meant to be that.
Have you guys ever worked in a real company?
Besides the idea ain't safe either: it's full of race conditions.
Whenever I release a new version of the product I'm working on, I prepare all anouncement mails beforehand. For reasons that are irrelevant here, these usually are not all identical. So I can't simply send one mail to an single list address. When all of them are ready (usually there are about 5), I manually send them out one right after the other. This will easily make me send more than 3 in less than a minute.
If I ever need to automate this procedure, my rate will be even higher.
It's not just users who are stupid. I know of an admin who actually used Goner to defend the ongoing introduction of Outlook at the place where he works.
Here's what happened: they were hit at 17:50 local time, at about 18:00, the first of four Outlook
lusers clicked on the attachment, which made the few admins who were still at work aware of the
problem. As they immediately went into action, they were able to get the mail servers under control pretty quickly (relatively speaking, that is). Next day, however, a scan of the network
revealed that about 50 additional PCs had to be cleaned up. These belonged to people who still use Netscape to read their mail and had also activated the worm. It didn't spread from there, but it did disable the virusscanners, so...
Next thing, that admin that I'm refering to claims: "Fortunately, we have Outlook installed on a few PCs already, because that is how we found out just before leaving for home. If everybody still used Netscape, a lot more PCs would have been infected during the evening, night, and morning before the helpdesk would have noticed the problem."
Going for the 20th anniversary edition is definitely worthwhile. It uses 20 years of hindsight to comment on what Brooks now thinks he was wrong about in the original edition (not a lot). Plus it adds his famous "No Silver Bullet" 1986 IFIPS paper.
Slashdot Mirror
Hey, I've done exactly that to a modern one not so long ago after opening up a keyboard in order to repair one key. It has worked fine ever since.
While I agree that the state of C++ compilers could be a lot better (more power to EDG!), the Mozilla C++ Portability Guide is seriously outdated. It is by now 4 years old, and it shows. The claims about the HP compiler, for instance, apply to the previous compiler. No, not just the previous version, the previous compiler, for which HP has discontinued all support some time ago already. Actually, even back in April 98, HP's CC compiler was already being phased out and replaced by the much much better aCC one. Actually, we made the switch some time (I don't recall exactly) before the start of my current project in September of 1997.
Whilst it is true that I think Microsoft is Vile, if it saves all of the threatened clubs my opinion of them will increase dramatically.
You don't really think MicroSoft (or any other big company for that matter) would buy the TV station to save the local football clubs, I hope... Big companies like that don't care at all about such things. As a matter of fact, neither do the vast majority of smaller ones.
To let your opinion of a company be influenced by this kind of "accidental" benefit is a recipe for serious disillusionement.
Looking at the site of my employer, browser statistics depend a lot on the content of the site, but apparently not as much as I thought.
We're a micro-electronics research centre (covering everything from MPEG-4 standard definition, over ECAD, chip packaging, down to our very own cleanroom). This means that we should have more Unix based visitors than the average me-and-my-cat home page. Our stats confirm his, but still MSIE takes between 80 and 90% (the figures vary on a daily basis), which is way more than I was expecting.
Fortunately, as of this week, we finally support all gecko based browsers. Hopefully, they will now start to rise above the disappointing 0.9% they got so far.
The problem with these fake browser strings, is that it helps to keep the Mozilla statistics down, thus giving the PHBs a reason why Mozilla support isn't a priority. That's exactly what I'm facing right now. Our web site doesn't work with anything except IE5+ and NS4 (due to crappy HTML). It took me a lot of lobbying to get people to work on fixing it at least partly, and even now I have to do part of it myself (it's not my job) because if they get stuck, chances are that they'll use the low percentage of Mozilla and Opera visitors to abort the action.
People in the know will know that fake entries exist, but those are so few and far between that it doesn't really matter. Besides, I doubt many admins look at individual entries. They most likely just look at automatically generated summaries. So I'd prefer all those who play games with this, to stop doing it except in exceptional circumstances.
If a site doesn't even let you in, it's not worth your visit. Our site is badly broken too at the moment, but it at least allows people with non-mainstream browsers to enter end get st part of the info.
Because opening up the plans removes the obscurity aspect without adding anything new to replace it. With open source, the white hats can actually fix the problems they see. With open building plans there's nothing Joe Ramdom Hacker can do to actually help improve security. So why would he care?
Also, building security is much easier and (most of all) after thousands of years of practice also much much better understood than software security.
All in all, in the building context, the risk that there are unknown flaws is much lower and the risk that only the black hats (who themselves happily operate under maximum obscurity) would benefit from an "open plan" is much larger.
Ever since so many years ago I had an eye infection and couldn't look at a lot of white light for hours on end, it is also exactly my default background color. At last I know why I've never gone back once the infection was cured: I have by pure coincidence stumbled onto nirwana. :-)
I tried to look at it with Mozilla 0.9.8 on Linux. The thing entered an endless loop, redisplaying the page over and over again. When I hit the back button, mozilla even crashed. So much for a working example site...
And before I forget: I'm actually using an HP-UX desktop (not for long anymore, but still). =Last time I looked, there still was no flash plugin for HP-UX. If Macromedia wants people to sort of drop HTML in favour of Flash, they better start by making their stuff available on ALL platforms, no matter how obscure.
I have a big pile of dead tree books about computing (in the broadest sense of the word) and do tend to look for more interesting ones whenever I have time (which doesn't happen a lot lately, however).
The one key characteristic of most of these books, is that they are (more or less) timeless. The kind of stuff that you can buy today and still use as a reference 5 or 10 years from now. In other words, I want books about concepts, not implementations.
I don't mind if any example code is written in a language that is slightly less hip when I'll be reading it than it used to be when it was written. Hell, I'll gladly buy an earlier edition that is 5 years old and uses what by now already are "all the wrong example systems" if I think the book has intrinsic value on the conceptual front and the new edition basically just translates the examples to something more modern.
What I'll almost never buy, is the kind of computer books that just teach you how to use today's (or rather yesterday's) version of program X or trendy API Y. Those I consider utter garbage. By the time you carried them home, they're hopelessly outdated.
I must admit to buying a few language books in the past, but even then nothing of the "Using Visual Sh.t version 3.1415" or the "Become a Java Pro in 30 Days" kind. To teach myself C++, for instance, I first read a copy of Stroustrup's second edition, which I got from a library, once. Linearly, that is. Then I started out for real by buying the ARM and the Coplien book. I bought a few more advanced books about C++ later, but basically those were not essential for me. By now, if I'd need to update my C++ knowledge to cover the STL more in depth, I'd "simply" dig up my aging copy of the draft standard.
Many thanks in advance for trying!
Indeed! xhpcalc is the best calculator program I have ever seen. By a long margin.
HP would make themselves very popular by releasing the source code for that thing, considering that they dropped it as long ago as the release of HP-UX 10.0 and that even so people still preserve it when installing new machines or OS versions.
Unfortunately, I suspect that xhpcalc literally emulates the calculators in question using their original code ROM images. That probably means it's very unlikely to ever see the light of day again, even if they can still find a backup tape containing the source.
Yes, it exists, we have it installed. But in a multi-platform environment, you want to support one browser and one browser only. Netscape is the only option in that case. Besides, you need HP-UX 11 to get a non-beta version of IE for HP-UX.
But all that was not my point. I can also use an NT PC if really needed. My point was that HP is giving a rather special message in doing things this way. A message that leaves me unwilling to believe some of the other stuff they claim.
I sure am hoping that this is not the way their Linux commitment will be implemented...
Europeans, ... cannot understand what real democracy is about.
Says an inhabitant of a country that doesn't even have something as basic a constitution, in which the head of state is not elected, and in which parties that get 40% of the votes of the people routinely end up with 60% of the seats in parliament.
I would suggest that you try thinking though the context of your statements before engaging in ignorant-Europe FUD.
Many more people understand and even speak multiple languages here in Europe than the average American can even start to imagine. And the percentage of people who don't do so is falling all the time.
Besides, it's not because people in Detroit and Dallas all speak American english that they move around just as easily as you imply. There are many more reasons than language to stick around in the same area.
I really wonder where you got that bit of FUD about the alloys. Everything is normed to be exactly the same, with the single exception of the back side of the coins. It simply has to be, otherwise vending machines would not work at all.
The only issues are whether all producers are capable of sticking to the allowed tolerances and whether those who programmed the vending machines didn't stick to closely to the official norm, ignoring (part of) the tolerances. The first issue should be a non-issue, I think. After all, it alos applied to the previous national coins. The second one can be corrected if needed.
That's because today's mechanic can easily fix yesteryear's car without requiring tons of additional training (finding spare parts might be a different thing though).
Another problem with this car analogy, is that old cars break over time as they get used. So in the end you need to replace them, thus allowing Honda & Co. to live. Old software doesn't break (remember, we're talking binaries here, not source code that is being patched up beyond recognition). It might be broken from day one, but if it was good enough for people to use anyway, it remains so until people's needs change. Hence, the software company needs an other way to keep the money flowing if it itself wants to stay alive.
This is why Microsoft is changing the licensing rules as of XP. Even they can't keep on bloating Word & Co. indefinitely. At some point nobody buys the new version anymore. Unless some licensing mechanism or a dependency on an external component forces them.
Good try, but it assumes that the mail sending program can be edited. This is not always the case.
I can't believe I'm reading this. And even less that someone moderated this up as being interesting. Sarcastic would have been more appropriate. Except that it likely wasn't meant to be that.
Have you guys ever worked in a real company?
Besides the idea ain't safe either: it's full of race conditions.
A lot of mail gets send fully automatically for very good reasons. That is what you are missing.
Are you for real?
Whenever I release a new version of the product I'm working on, I prepare all anouncement mails beforehand. For reasons that are irrelevant here, these usually are not all identical. So I can't simply send one mail to an single list address. When all of them are ready (usually there are about 5), I manually send them out one right after the other. This will easily make me send more than 3 in less than a minute.
If I ever need to automate this procedure, my rate will be even higher.
Telling them that their e-mail access will be cut off does not help if they are the CxO (for some random x). Of course, if it's the CIO ...
It's not just users who are stupid. I know of an admin who actually used Goner to defend the ongoing introduction of Outlook at the place where he works.
Here's what happened: they were hit at 17:50 local time, at about 18:00, the first of four Outlook
lusers clicked on the attachment, which made the few admins who were still at work aware of the
problem. As they immediately went into action, they were able to get the mail servers under control pretty quickly (relatively speaking, that is). Next day, however, a scan of the network
revealed that about 50 additional PCs had to be cleaned up. These belonged to people who still use Netscape to read their mail and had also activated the worm. It didn't spread from there, but it did disable the virusscanners, so...
Next thing, that admin that I'm refering to claims: "Fortunately, we have Outlook installed on a few PCs already, because that is how we found out just before leaving for home. If everybody still used Netscape, a lot more PCs would have been infected during the evening, night, and morning before the helpdesk would have noticed the problem."
Sadly, this really is a true story...
Going for the 20th anniversary edition is definitely worthwhile. It uses 20 years of hindsight to comment on what Brooks now thinks he was wrong about in the original edition (not a lot). Plus it adds his famous "No Silver Bullet" 1986 IFIPS paper.