and of course, you can always purchase Ricochet transport to make up for the lost point-to-point link...and one could also argue that micro-cellular networks are more bandwidth efficient than long-haul point-to-point links...but then I'd have to duck!
Now you can only get Internet service through Ricochet -- all devices, subscribed after December 23 2000, are not allowed to talk anything but ISPs' access point. I had to go through a lot of bullshit before Ricochet finally confirmed that -- their network supports "peer to peer" links, but they decided to disable it "because they don't want to support it".
Great -- I have "upgraded" Ricochet GS to Merlin for Ricochet, and now it can't talk to my home backup link (another Ricochet GS "modem"), so when WWC has an outage on its (single?) router, new shiny Merlin for Ricochet is for all practical purposes indistinguishable from pokemon card stuck into PCMCIA slot.
If people get to make their own decisions (which is the case in a free market), then how is someone able to force their decisions on everybody else? Either you didn't write what you mean, or what you mean is incomprehensible.
No sane person will make "his own decisions" when confronted on any important issue (more important than, say, price of a bagel in a coffee shop) -- he will look for someone else to make sure that large enough number of participants will make the same decision, so it will either become the only solution available to the opponent, or at least represent large enough piece of the market.
This is what I,mean by "negotiating power", and this is what in simple case of prices and salaries companies exercise when they standardize their products and what workers exercise when they join unions. When things are more important than money (health, civil liberties, life), even if conflict doesn't lead to direct physical confrontations, various groups of people and organizations start all kinds of exercises of negotiating power -- recent example is Ashcroft's confirmation, where more organized Republicans forced large number of Democrats to vote contrary to their true opinion.
In the case of privacy the conflict is over a liberty of being able to keep private information. People want to keep it, however they need very complex and costly process of organizing a boycott to force companies to change their policies -- that forces each person to act as if very little of negotiating power is available to him. OTOH, companies have no trouble of creating all kinds of groups and alliances that force uniform privacy standards -- companies' executives expect that whatever horrendous infringemnet of personal privacy will be in the standard, joining the group will be more beneficial to the company than establishing a better privacy standard, as customers will expect low privacy, and once lost their personal information to one of companies in the group, they won't see any benefit in not "losing it again" to another company. To make things worse, companies may choose monopolists with some essential service (say, FedEx and UPS) and offer large payments for disclosure of customers' data to the rest of the group, thus making any attempt by other companies to respect privacy absolutely pointless, even if the original anti-privacy group will be small.
"Let's declare that something can be traded on a free market, and everything will automagically adjust itself".
Things don't work this way -- if a decision how to value something is left for everyone in each case, ones with more negotiating power, force their decisions on everybody else. In this case corporations who will create their (low) privacy standards will easily leave consumers with no choice, and consumers would have to resort to inefficient and extremely hard to organize boycotts to get anything back.
I have seen people who honestly believed in Communist utopia, and they made more sense than people who honestly believe in this Libertarian utopia.
IE & standards: Up to NS6, IE 5.5 (& IE 5 for Mac) was the most standards complaint browser. (If this doesn't mean much to, this mean that it supported standards very well.)
Netscape has a tendecy to freak at non-100%-correct(-to-netscape's-standards) HTML.
I don't know about you, but while I think that 100%-correct is a Good Thing, I don't think that the browser should *force* it on you.
Not everybody can be bothered learn HTML, and WYSIWYG isn't perfect either.
All MSIE does is guessing, what author could mean -- it's impossible to parse correctly all the crap that displays in MSIE without wasting huge amount of time and effort on doing it. While Mozilla supports most of broken HTML, it's impossible to make it work exactly like MSIE does without having precisely the same parsing+guessing algorithm -- precisely because that broken HTML wouldn't be broken if it was not ambiguous.
If I write a page for IE4, I can be reasonably sure that it would display correctly on IE4+ (that is why most people use document.all for IE, instead of the standard way, btw)
If you are so dumb that you can't write a page that complies with standards and doesn't trigger known bugs/incompatibilities in IE and Netscape, you write a page "for IE4+", and this is why it doesn't work everywhere else. But that makes you a moron.
Mozilla & memory: I've tested NS6 on 2k machine.
I neither know nor care how NS6 works on W2K. Mozilla.7 on Linux takes 25-30M, and I don't think that MSIE with all its DLLs uses less. I still would prefer a better browser, but MSIE isn't noticeably better.
Sorry, but they didn't screwed up Netscape.
Netscape did it to itself all by its own.
It started with total ignorance of standards, went on to *two* years (what is that in internet years?) with no meaningful update. It was buggy as hell, heavy, and unfriendly.
IE has little to do with standards, too -- most of pages that don't display properly in Netscape but display in IE, are complete bullshit from any set of standards' point of view. It's possible to make standards-compliant page that doesn't work in Netscape, but in reality I have yet to see it (not to mention that MSIE by itself supports standards poorly, too -- a lot of things will display in Netscape or Mozilla, but won't in MSIE).
And it gets worse if you open several windows at the same time, I've opened three NS6 windows and watched as it ate 65MB. I currently have 5 IE windows open, and IE takes 9MB (peaked at 20MB)
I sometimes has up to 20 - 25 windows open, I don't feel like buying Gigabytes of RAM just to accomedate NS.
You counted your memory wrong. Mozilla (that you without any doubt ran on Linux) is multithreaded, so all those processes that you have seen occupy the same memory.
SOAP ties together Bonobo, CORBA, DCOM, and any other object type you want.
So does TCP. Without information about meaning of the data in XML it's as proprietary as any binary, and DTDs or schemas do nothing to describe how to operate on the data. Also "remote objects" without objects mirroring mechanism and transparent transfer, are only good as a recipe for network congestion -- if you think, X is slow, try to imagine how bad that thing will be.
Really, all that ".NET" is a marketing campaign for DCOM, with protocol changed from whatever bullshit it used before to SOAP. What basically is yet another RPC.
So far all technologies incorporated into.NET except COM and XML, are complete failures - RPC in all of its incarnations is most hated protocol ever, DCOM is a bitch to write for, and ActiveX never took off (except as calling anything that uses COM "ActiveX"). COM wasn't a complete failure because Microsoft was pushing it for more than a decade, and XML is a "technology" in the same way as comma-separated list is a "technology".
Copyright laws aren't mutating into anything -- there was no changes in copyright laws except some ridiculous extension of the duration of protection. DMCA is a kind of legal garbage that has nothing to do with "copyright laws" -- it's an exception from general laws made for particular kind of distributors, legalizing practices that amount to extortion and illegal terms of contract.
...because all "copyright protection" mechanisms give the distributor full control over everything recorded -- what is a kind of "copy protection" that is infinitely higher than anything allowed to him by copyright laws. It's a way to strip users from any rights, protected as fair use, and put content distributor in charge of policing itself.
No possible technical "copyright protection" measure can't be reduced to limiting what distributor can do, therefore all of them -- claiming to be "copy protection", "copyright protection" or anything else, should be fought against.
I can write a config and script for plugger to run vmware from in inside Netscape/Mozilla/..., and boot in it from downloaded image for any OS that runs on x86. With all secuity and whatnot -- will I get a slashdot story for that?
Java merely conceals memory leaks by converting them to almost undetectable reference leaks -- different ways, same result, as after few hours or days of running program runs out of memory and crashes (usually crashing a bunch of other stuff in the process). I'll rather have an immediately detectable dangling pointer or traceable leak in C or C++.
... you must make your code more simple, as you will never catch all possible troubles with a debugger. If multithreading caused it to be more complex, switch to either single thread with nonblocking operations, multiple processes or any combination of two.
And, of course, Java will ALWAYS remain inadequate by both performance and reliability -- five years is more than enough to fix a bad product even for Sun (ex: Solaris) if it really cared about doing that.
GTK is very fast. However I had to change $HOME to make Netscape 6 ignore my GTK theme that was absolutely useless for Netscape (it has XUL-defined interface that overrides everything) but for some reason still was read by it.
Another thing that affects the performance is memory -- since Windoze-using people that also have Linux box usually have at least twice more RAM in Windows box, Netscape 6 with its large memory footprint will be slower on their Linux box.
I've seen all kinds of numbers, from IE having 98% of the browser share, to Netscape having 10% of the browser share. Any idea what Netscape's browser share is like, especially with the release of Netscape 6?
Windows + MacOS is installed on much less than 98% of desktop boxes, so MSIE market share theoretically can't be 98%.
Yes. And I didn't forget to install truetype fonts, so it works perfectly with documents that were written using the same fonts by Microsoft lamers like you (Microsoft formats are so shitty, minimal change in font totally screws up the layout, so if you run StarOffice with minimal set of X fonts, most of documents will be be barely recognizable -- but then, I dare you to run Word on Windows with deleted Arial and Times New Roman fonts).
How many dtp people use Linux? These people are not, in the main, technical types into kernel recompilation. So no market for the product. Simple economics - I have graphics dudes around me, and they like Apple Macs, and, in a few cases, Windows - they don't like Unix, and couldn't use it.
What DTP? Framemaker is also used for a lot of technical documentation -- something that developers often write or edit on their boxes. Considering that Linux is now replacing Solaris as "average Unix" development platform in many places, it's very reasonable to have Framemaker on it.
Yeah. But you know English. What about those Russians who know no English? And I guess they aren't a minority. Why can't they participate in a Russian sub set of Internet?
They can, and do. It doesn't take a knowledge of English to type in a domain name -- especially if the name is actually a Russian word in transliteration. I started working with computers when I had the same amount of English knowledge as most of Russians did in 1986 (almost none) and still I had no problems typing commands and program names in English, as long as I could read and edit text in Russian. I think, it was 1990 when I actually became able to read English text more or less easily, and at that time I already completed few large software projects.
There are a lot of things where having material in Russian helps a lot, but domain names (just like program names) isn't one of them.
I neither know nor care, why Chinese government is doing this, but the whole idea of Unicode in DNS is stupid, counterproductive and serves no purpose other than more money for somain registrars and software manufacturers that will all issue "compulsory" upgrade versions of their software to support it.
Oh, BTW, I am Russian, and the last thing I need is Russian in domain names, especially on computers that have no russian keyboard.
Yes, they have moved web servers to Windows 2000. However they didn't mention in their announcements that those servers are behind TWO LAYERS OF CISCO LOCAL DIRECTORS. I have done some probing of that monster, and the only possible reason for this is huge number of boxes where load is supposed to be distributed. The funniest thing is, any response from www.hotmail.com that normal browsers receive (that should be a redirect to one of Local Directors on the second layer) comes directly from Local Director on the first layer and not a single byte passes through Windows box. There is a way to get to the Windoews box behind it, but one has to put HTTP request into multiple packets to get a response from there.
First layer consists of Local Directors on every IP address that corresponds to www.hotmail.com, they only send HTTP redirects to the second layer (violating HTTP protocol in the process). Second layer has actual servers behind them, and what happens next is hidden from the HTTP clients, so I can only guess how huge and inefficient is the rest of the system. However if someone wants to bring Hotmail down, he only has to overload Local Directors, and being relatively dumb, Cisco Local Directors should be easy to convince that few thousands of packets sent from single box are in fact thousands of users that simultaneously sent a bunch of HTTP requests.
If they need Exchange-like features other than calendar (shared message folders, etc.), just install cyrus. Probably a good calendar server exists, too, but having no need in such a thing I never looked.
Slashdot Mirror
and of course, you can always purchase Ricochet transport to make up for the lost point-to-point link...and one could also argue that micro-cellular networks are more bandwidth efficient than long-haul point-to-point links...but then I'd have to duck!
Now you can only get Internet service through Ricochet -- all devices, subscribed after December 23 2000, are not allowed to talk anything but ISPs' access point. I had to go through a lot of bullshit before Ricochet finally confirmed that -- their network supports "peer to peer" links, but they decided to disable it "because they don't want to support it".
Great -- I have "upgraded" Ricochet GS to Merlin for Ricochet, and now it can't talk to my home backup link (another Ricochet GS "modem"), so when WWC has an outage on its (single?) router, new shiny Merlin for Ricochet is for all practical purposes indistinguishable from pokemon card stuck into PCMCIA slot.
Ricochet uses 900MHz, both old and new systems.
If people get to make their own decisions (which is the case in a free market), then how is someone able to force their decisions on everybody else? Either you didn't write what you mean, or what you mean is incomprehensible.
No sane person will make "his own decisions" when confronted on any important issue (more important than, say, price of a bagel in a coffee shop) -- he will look for someone else to make sure that large enough number of participants will make the same decision, so it will either become the only solution available to the opponent, or at least represent large enough piece of the market.
This is what I ,mean by "negotiating power", and this is what in simple case of prices and salaries companies exercise when they standardize their products and what workers exercise when they join unions. When things are more important than money (health, civil liberties, life), even if conflict doesn't lead to direct physical confrontations, various groups of people and organizations start all kinds of exercises of negotiating power -- recent example is Ashcroft's confirmation, where more organized Republicans forced large number of Democrats to vote contrary to their true opinion.
In the case of privacy the conflict is over a liberty of being able to keep private information. People want to keep it, however they need very complex and costly process of organizing a boycott to force companies to change their policies -- that forces each person to act as if very little of negotiating power is available to him. OTOH, companies have no trouble of creating all kinds of groups and alliances that force uniform privacy standards -- companies' executives expect that whatever horrendous infringemnet of personal privacy will be in the standard, joining the group will be more beneficial to the company than establishing a better privacy standard, as customers will expect low privacy, and once lost their personal information to one of companies in the group, they won't see any benefit in not "losing it again" to another company. To make things worse, companies may choose monopolists with some essential service (say, FedEx and UPS) and offer large payments for disclosure of customers' data to the rest of the group, thus making any attempt by other companies to respect privacy absolutely pointless, even if the original anti-privacy group will be small.
"Let's declare that something can be traded on a free market, and everything will automagically adjust itself".
Things don't work this way -- if a decision how to value something is left for everyone in each case, ones with more negotiating power, force their decisions on everybody else. In this case corporations who will create their (low) privacy standards will easily leave consumers with no choice, and consumers would have to resort to inefficient and extremely hard to organize boycotts to get anything back.
I have seen people who honestly believed in Communist utopia, and they made more sense than people who honestly believe in this Libertarian utopia.
IE & standards: Up to NS6, IE 5.5 (& IE 5 for Mac) was the most standards complaint browser. (If this doesn't mean much to, this mean that it supported standards very well.) Netscape has a tendecy to freak at non-100%-correct(-to-netscape's-standards) HTML. I don't know about you, but while I think that 100%-correct is a Good Thing, I don't think that the browser should *force* it on you. Not everybody can be bothered learn HTML, and WYSIWYG isn't perfect either.
All MSIE does is guessing, what author could mean -- it's impossible to parse correctly all the crap that displays in MSIE without wasting huge amount of time and effort on doing it. While Mozilla supports most of broken HTML, it's impossible to make it work exactly like MSIE does without having precisely the same parsing+guessing algorithm -- precisely because that broken HTML wouldn't be broken if it was not ambiguous.
If I write a page for IE4, I can be reasonably sure that it would display correctly on IE4+ (that is why most people use document.all for IE, instead of the standard way, btw)
If you are so dumb that you can't write a page that complies with standards and doesn't trigger known bugs/incompatibilities in IE and Netscape, you write a page "for IE4+", and this is why it doesn't work everywhere else. But that makes you a moron.
Mozilla & memory: I've tested NS6 on 2k machine.
I neither know nor care how NS6 works on W2K. Mozilla .7 on Linux takes 25-30M, and I don't think that MSIE with all its DLLs uses less. I still would prefer a better browser, but MSIE isn't noticeably better.
Sorry, but they didn't screwed up Netscape. Netscape did it to itself all by its own. It started with total ignorance of standards, went on to *two* years (what is that in internet years?) with no meaningful update. It was buggy as hell, heavy, and unfriendly.
IE has little to do with standards, too -- most of pages that don't display properly in Netscape but display in IE, are complete bullshit from any set of standards' point of view. It's possible to make standards-compliant page that doesn't work in Netscape, but in reality I have yet to see it (not to mention that MSIE by itself supports standards poorly, too -- a lot of things will display in Netscape or Mozilla, but won't in MSIE).
And it gets worse if you open several windows at the same time, I've opened three NS6 windows and watched as it ate 65MB. I currently have 5 IE windows open, and IE takes 9MB (peaked at 20MB) I sometimes has up to 20 - 25 windows open, I don't feel like buying Gigabytes of RAM just to accomedate NS.
You counted your memory wrong. Mozilla (that you without any doubt ran on Linux) is multithreaded, so all those processes that you have seen occupy the same memory.SOAP ties together Bonobo, CORBA, DCOM, and any other object type you want.
So does TCP. Without information about meaning of the data in XML it's as proprietary as any binary, and DTDs or schemas do nothing to describe how to operate on the data. Also "remote objects" without objects mirroring mechanism and transparent transfer, are only good as a recipe for network congestion -- if you think, X is slow, try to imagine how bad that thing will be.
Really, all that ".NET" is a marketing campaign for DCOM, with protocol changed from whatever bullshit it used before to SOAP. What basically is yet another RPC. So far all technologies incorporated into .NET except COM and XML, are complete failures - RPC in all of its incarnations is most hated protocol ever, DCOM is a bitch to write for, and ActiveX never took off (except as calling anything that uses COM "ActiveX"). COM wasn't a complete failure because Microsoft was pushing it for more than a decade, and XML is a "technology" in the same way as comma-separated list is a "technology".
Copyright laws aren't mutating into anything -- there was no changes in copyright laws except some ridiculous extension of the duration of protection. DMCA is a kind of legal garbage that has nothing to do with "copyright laws" -- it's an exception from general laws made for particular kind of distributors, legalizing practices that amount to extortion and illegal terms of contract.
...because all "copyright protection" mechanisms give the distributor full control over everything recorded -- what is a kind of "copy protection" that is infinitely higher than anything allowed to him by copyright laws. It's a way to strip users from any rights, protected as fair use, and put content distributor in charge of policing itself.
No possible technical "copyright protection" measure can't be reduced to limiting what distributor can do, therefore all of them -- claiming to be "copy protection", "copyright protection" or anything else, should be fought against.
I can write a config and script for plugger to run vmware from in inside Netscape/Mozilla/..., and boot in it from downloaded image for any OS that runs on x86. With all secuity and whatnot -- will I get a slashdot story for that?
Java merely conceals memory leaks by converting them to almost undetectable reference leaks -- different ways, same result, as after few hours or days of running program runs out of memory and crashes (usually crashing a bunch of other stuff in the process). I'll rather have an immediately detectable dangling pointer or traceable leak in C or C++.
... you must make your code more simple, as you will never catch all possible troubles with a debugger. If multithreading caused it to be more complex, switch to either single thread with nonblocking operations, multiple processes or any combination of two.
And, of course, Java will ALWAYS remain inadequate by both performance and reliability -- five years is more than enough to fix a bad product even for Sun (ex: Solaris) if it really cared about doing that.
GTK is very fast. However I had to change $HOME to make Netscape 6 ignore my GTK theme that was absolutely useless for Netscape (it has XUL-defined interface that overrides everything) but for some reason still was read by it. Another thing that affects the performance is memory -- since Windoze-using people that also have Linux box usually have at least twice more RAM in Windows box, Netscape 6 with its large memory footprint will be slower on their Linux box.
I've seen all kinds of numbers, from IE having 98% of the browser share, to Netscape having 10% of the browser share. Any idea what Netscape's browser share is like, especially with the release of Netscape 6?
Windows + MacOS is installed on much less than 98% of desktop boxes, so MSIE market share theoretically can't be 98%.
Have you used Star Orifice??
Yes. And I didn't forget to install truetype fonts, so it works perfectly with documents that were written using the same fonts by Microsoft lamers like you (Microsoft formats are so shitty, minimal change in font totally screws up the layout, so if you run StarOffice with minimal set of X fonts, most of documents will be be barely recognizable -- but then, I dare you to run Word on Windows with deleted Arial and Times New Roman fonts).
How many dtp people use Linux? These people are not, in the main, technical types into kernel recompilation. So no market for the product. Simple economics - I have graphics dudes around me, and they like Apple Macs, and, in a few cases, Windows - they don't like Unix, and couldn't use it.
What DTP? Framemaker is also used for a lot of technical documentation -- something that developers often write or edit on their boxes. Considering that Linux is now replacing Solaris as "average Unix" development platform in many places, it's very reasonable to have Framemaker on it.
What you have seen was an "access denied" message.
Yeah. But you know English. What about those Russians who know no English? And I guess they aren't a minority. Why can't they participate in a Russian sub set of Internet?
They can, and do. It doesn't take a knowledge of English to type in a domain name -- especially if the name is actually a Russian word in transliteration. I started working with computers when I had the same amount of English knowledge as most of Russians did in 1986 (almost none) and still I had no problems typing commands and program names in English, as long as I could read and edit text in Russian. I think, it was 1990 when I actually became able to read English text more or less easily, and at that time I already completed few large software projects.
There are a lot of things where having material in Russian helps a lot, but domain names (just like program names) isn't one of them.I neither know nor care, why Chinese government is doing this, but the whole idea of Unicode in DNS is stupid, counterproductive and serves no purpose other than more money for somain registrars and software manufacturers that will all issue "compulsory" upgrade versions of their software to support it.
Oh, BTW, I am Russian, and the last thing I need is Russian in domain names, especially on computers that have no russian keyboard.
See http://slashdot. org/com ments.pl?sid=00/11/16/194209&cid=737. What Netcraft sees is a redirect header, made by Cisco Local Director -- not a single byte in it came from Windows box.
Yes, they have moved web servers to Windows 2000. However they didn't mention in their announcements that those servers are behind TWO LAYERS OF CISCO LOCAL DIRECTORS. I have done some probing of that monster, and the only possible reason for this is huge number of boxes where load is supposed to be distributed. The funniest thing is, any response from www.hotmail.com that normal browsers receive (that should be a redirect to one of Local Directors on the second layer) comes directly from Local Director on the first layer and not a single byte passes through Windows box. There is a way to get to the Windoews box behind it, but one has to put HTTP request into multiple packets to get a response from there. First layer consists of Local Directors on every IP address that corresponds to www.hotmail.com, they only send HTTP redirects to the second layer (violating HTTP protocol in the process). Second layer has actual servers behind them, and what happens next is hidden from the HTTP clients, so I can only guess how huge and inefficient is the rest of the system. However if someone wants to bring Hotmail down, he only has to overload Local Directors, and being relatively dumb, Cisco Local Directors should be easy to convince that few thousands of packets sent from single box are in fact thousands of users that simultaneously sent a bunch of HTTP requests.
I don't know about Outlook, but IE port to Unix never worked properly and no longer being developed.
If they need Exchange-like features other than calendar (shared message folders, etc.), just install cyrus. Probably a good calendar server exists, too, but having no need in such a thing I never looked.
Will the fungus be allowed to vote?
I don't know, however since some Florida voters behaved pretty much like fungus, why not.