SIP is the emerging standard for things like Internet telephony and is designed for general-purpose session negotiation. Windows (MSN) Messenger supports this today. The protocol supports basically anything you could need in an IM application and it's still flexible enough that vendors can still build on nifty extensions while maintaining compatibility with any other SIP client.
To be fair, while I've never been in one of his tutorial sessions, I've found mjd to be a very bright guy and a very entertaining speaker. His presentations at YAPC were heavily attended, and I found them interesting and easy to follow, especially this one.
Would you rather he not have spoken up before the tutorial began, in this case? Let people throw their $800 away without knowing that they already know the material until it's far too late?
Regarding the lack of multi-window support, I agree. I've played with XFree86 for Win32 quite a lot, and this is the chief and only significant reason it's unsuitable for the things I do.
With everything in one big X window, I *have* to load an X window manager and have to keep all of my X apps contained in that one window. It makes it a little awkward if I need to be working between a Windows app and an X app, because I'm having to swap out of the X system entirely to go back to a Windows app. It's just a little cumbersome. I love the way Exceed is able to do this: every X app appears seamlessly like any other Windows app. And it's fast.
Plus, it's sometimes interesting to still use a window manager or shell in this mode. You get the benefits of having X windows float around a Win32 desktop, in addition to having these windows all managed by a different window manager. It's kind of cool having your Gnome panel, desktop icons, etc., floating on your Win32 desktop alongside your Explorer bar and Explorer desktop icons.
I disagree. If you're sticking something up on a web site, that something has a URL. Every entity on a web site has its own unique URL that should be retrievable anywhere.
If you don't like this behavior, and you want "pages" on your site to only be accessible by people browsing through your site, you're going to need to stick a "document retrieval" application layer onto your site. Users start a session when they enter this application, and are only able to retrieve stories through this application front-end. This can be done through HTTP as simply as with a session ID, but the web was not meant to work like this.
Again, we have a rather useful technology being twisted and warped by corporate interests instead of those corporate interests funding a proper technological solution, just like the intellectual property crap associated with DNS nowadays.
Other benefits of normalizing on a distance based on the speed of light is that c can become 1, greatly simplifying things like e=mc^2 to e=m^2, for instance. We can essentially just start eliminating the speed of light from equations entirely, since it will (generally) just factor out if it's equal to one.
We have a large number of units that depend directly upon the unit we use for time. Something fundamental as the Hertz would be redefined if we choose to use something new for a "metric" second (ignoring the fact that a "second" is already an SI unit).
The biggest problems with our method of measuring time (the requirement for leap years, leap seconds, etc.) won't be solved just by changing systems, since our planet doesn't rotate at a perfect rate and doesn't orbit the sun at an integral number of days.
You have to base a new system upon the rotation of the earth, because any time system that doesn't make sense relative to a single earth day would not be accepted. So while we can very nicely subdivide a day, a "day" is still pretty arbitrary and specific to the earth. When we deal with astronomical events, or have to tell time on another planet, we're still stuck with doing it like they do on earth.
It might make more sense from a scientific point of view to adopt a "second" that's more closely tied in "metric" respects to the half-life of a known element, or perhaps something like the absorption spectrum of hydrogen. (Normalizing to the latter, keeping the base unit as close as possible to a traditional second, a unit equivalent to exactly 1 billion times the absorption frequency of hydrogen would be about 0.7 traditional seconds, which would put closer to 123000 of them in a day.)
Units of distance should be measured (regardless of our definition of "second") in some form of light-seconds. Using our new and improved second, 1 nano-light-second would be equivalent to about 21cm (or nearly 8-1/2 inches). Not coincidentally, this also happens to be the wavelength of hydrogen absorption. Not too cumbersome, eh?
Some states do have statutes prohibiting caller ID blocking for telemarketing calls. Unfortunately a lot of the telemarketing calls I receive are simply "OUT OF AREA" and are not blocked, just unavailable. Since I get one or two other regular people that are "OUT OF AREA" calling me, I have to pick these up as well.
Fortunately I'm on my state's no-call list, so those that do call me get an earfull.
Agreed, even in my case where I have two monitors on a WinXP system, I have more than enough space to keep all of my background apps running *and* still be productive with Photoshop or some other big app.
However, I do also have a third cheap monitor off to the side dedicated to receiving syslog messages and other text bits of importance from my Linux systems. It might be nice to offload some of this onto an LCD screen that I can use a little more portably. It's not my screen's real-estate that I'm concerned about, it's my desk's!
Why totally impractical? Use 'ruptime' to gather the stats, and a stock LCD app to send the data to the display..
This is exactly the problem
on
ICANN Updates
·
· Score: 1
"good, common words domain names can generate hits on their own."
This is why we're in the mess we're in. Hostnames and URL's were meant for a time when the Internet was in the hands of the techies. When business and legal interests completely overwhelmed the purpose of the Internet, nothing better stepped in to supplement the current naming system.
We basically need a new directory service: one that maps legal entity names to DNS domains. Legislate this with all of the intellectual property crap you want, but leave DNS itself out of it. There's no reason users need to really see DNS domain names, and little reason they'd even need to see URL's for that matter. What percentage of the general public understands http:// and ftp://?
I should be able to tell my Internet browser to go to "IBM", and it should be able to consult a directory of entities, find a list of legal entities that have the name "IBM", obtain a DNS domain name from that directory (ibm.com), do an SRV lookup for the HTTP service at ibm.com, make the request for / at this server, and boom, you're at the "IBM" organization's home page. A TLS/SSL certificate at this point could validate the organization's identity (it should match the original directory entry), for those looking at security. Nowhere in this process do I have to guess that www.ibm.com is IBM's home page. The hostname in this case ceases to be such a valued commodity, outside of vanity uses.
We no longer see companies registering a hundred 2nd-level domains for every service mark and product they sell. For that type of thing, RealNames or similar services would (and do) work fine.
Re:Repeating myself
on
ICANN Updates
·
· Score: 3, Insightful
DNS names were never intended to be commodities (at least not to the degree they are today), any more than Java class name hierarchies or SNMP MIB organizational identifiers are meant to be commodities. DNS names have become commodities because there is presently no directory that associates legal business names to DNS domains. DNS is being warped into serving two purposes at once here: the association of names to IP addresses, and the association of legal business entities to an IP address for a web site.
What we need is to return DNS to serving its original role: to provide an easier-to-use addressing mechanism for Internet hosts. The role of associating legal entity names to Internet domains needs its own service (e.g. X.500 or LDAP). A "keyword" lookup service for product names or other service marks would need a third service a la RealNames.
We need to desperately curb the use of www.what-i-am-looking-for.com and to start enforcing DNS delegation like it was originally designed.
At first, before reading this into more detail, I was on the side of the Slashdot masses here. It seemed unreasonable that they'd be unwilling to share the list. What have they got to hide?
But then, after actually reading the decision, it made sense to me. By publishing the list, they would be basically providing to the world a nice, comprehensive list of, among other things, child porn sites, nicely categorized.
Now granted, child porn probably does not comprise the entirety of this list, but is it really worth turning the Australian government into the world's best child porn links site just to add your own level of oversight into their blocking?
I might suggest a couple of slightly different (non-mutually-exclusive) approaches to handling this trust issue:
1. Create a separate organization specifically to act as oversight for the list. Put them in a position to have access to this list, but keep the list itself out of general circulation. If you don't trust the government, maybe you can put some people you do trust into this oversight panel.
2. In addition to categorizing links to content as safe or unsafe, categorize the link itself along with it. If the group doesn't feel that link needs to be distributed, it can be excluded from general distribution (this solving the government's concerns over release). Everything else would make a publicly-available list and can enjoy all of the public scrutiny you like. This doesn't help if you believe the censors will be abusing their power by censoring things they shouldn't (since they'd just keep the link from you as well), but it would aid in catching accidental classifications, or where the classifications are likely to be contested.
Don't get me wrong, though, I don't really approve these acts of censorship either. But if the Australian government has decided to do it, they're going to do it. We can work to help them do it right without hindering the movement to keep them from doing it at all.
In addition, it should be fairly easy to identify censorship when you see it. To date I don't think I've read anything about how their censorship program has blocked something that shouldn't be. Maybe we can give them the benefit of the doubt (for now)?
And this would be different from RealNames because RealNames would only allow the association of that one word to that one company, so it suffers some of the same problems as DNS.
A proper directory service wouldn't care. If your business's name is Yahoo, that would go into the directory in the proper place, and a search of the directory would give you what you're looking for. It's not the convenience of 1-word "keywords" in some cases, but it beats having to guess a hostname and, thus, giving hostnames a monetary value and a herd of companies and lawyers ready to exploit that.
I'm kind of curious about your line of reasoning here. I think DNS as it's designed does a perfectly adequate job at resolving names.
I think it's a good thing that only "A" records take you from a name to an IP address. All other record types should only permit a name. Administration would be annoying otherwise. Same with CNAME records. If I want to express the fact that www.company-name.com should be the same as www.company.com, and we're in different groups, we shouldn't have to coordinate IP address changes. I just set up a CNAME record telling everyone my "www" is really the other "www", and to go get the "A" mapping from there. (On a side note, I completely disapprove of using this method to alias one web server to the other, unless the destination web server issues a redirect to the canonical web site URL.)
DNS is for providing information about domain names, including (but not limited to) IP addresses. You also have service information (e.g. MX and SRV records), contact information, etc. Any piece of data you need to associate with a DNS domain name should be expressible with the DNS protocol.
DNS as it is does its job fine, I have no problem with that. It's the fact that we're trying to squeeze intellectual property status to these assigned names that's the problem.
It's no more immense than DNS (which, admittedly, is pretty freakin' immense). X.500 (and LDAP) is designed to be distributed. You'd have a root (like DNS), but no "generic" top-level domains. You'd just hang each country's X.500 directory off of the root, and each country would be responsible for getting its local organizations listed in its directory. For the US, this could mean sub-delegations at the state level, sort of like how the.us DNS domain works. Traffic between tiers should be authenticated, a la DNSSEC (but who uses that?). When a business (or individual, for that matter) gets registered, they go into their local X.500 (or LDAP) directory and are immediately visible to the globe. This has the added benefit of doubling as a global telephone or address directory, with whatever information the user wishes to publish.
But really I'm kind of digressing. This is just one potential method for accomplishing the same goal: reducing DNS back to what it was designed for and creating something else to fill the demand for a look-up service, designed for that purpose.
Somebody else mentioned RealNames, and that's a step in the right direction, but it's still trying to use the "one word" approach to resolving trademarks. Why should 'Apple Computer' have any more claim over 'Apple' than 'Apple Hardware Store'?
Expanding upon this comment, the directory service should not even be web-centric. It should just associate a legal business name with the assigned DNS domain name (and/or telephone number, physical address, whatever). Using DNS SRV records, you could then associate a web server or an internal directory server with this domain name. All of this can be done without having to guess that "www.company-name.com" is the home page for a company. You'd just ask for "Joe's Widgets", and you'd get back joes-widgets.co.yz.us (or whatever). This domain would have an SRV record pointing to perhaps web.it.joes-widgets.co.yz.us, which would be a web server located in Joe's Widget's IT group. A certificate presented by the server would authenticate it as Joe's Widgets of Anytown, YZ, as certified by the state of YZ, as certified by the US, and your location effort would be complete.
But search engines as they are today have one flaw: it's difficult to find authoritative information.
If you want to find organizational information (e.g. a web site) for, say, Joe's Widgets located in Anytown, YZ, you might do a netsearch like you imply. This netsearch might come back with one or more web pages related to this company, and if you're lucky, the first link might even be the home page for this company itself.
But realistically, we don't have a "phone book" equivalent for the Internet. That's what a real directory service can provide that search engines, today, do not.
I'm not saying that commercialization of the Internet should not have occurred, I just think that steps should have been taken early on to allow that to happen in a more sane fashion. Treating the DNS hierarchy as an intellectual property battle-ground is silly. DNS domains should have exactly the same weight as any assigned name or number.
What are we going to see next? IP battles for control over different Java class name hierarchies? What about SNMP MIB organizational identifiers?
DNS is a technological tool, and should not be a business tool (like it's become).
DNS needs to be replaced
on
Ruling the Root
·
· Score: 4, Interesting
I'm interested in reading the book's take on this, but it seems to me that DNS was never meant to be twisted into what it's being used for today.
DNS's primary purpose was to allow someone to lookup a numerical address from a hierarchial alphanumeric one. It was never meant to be a search mechanism for people to find an organization. That's what directories like X.500 and LDAP, or search engines are for. The very fact that there can be multiple commercial entities primarily identified with the same word makes DNS wholly unsuitable for this.
I think we need to scrap DNS and start over from scratch. Hand out names as needed for technological reasons, and set up a directory for users to provide an organizational name and get back links to that organization's web site(s). Regulate that directory with all of the intellectual property and trademark crap that you want, but do it in such a way that local IP and trademark law can be respected. Hide DNS (and maybe URL's entirely for that matter) from the user through their browsers, and utilize certificates for establishing identity, instead of the basic assumption that if the domain name in the URL "looks right", then it must be them. I might even go so far as to suggest that when an organizational's DNS domain is created, they automatically get a certificate that allows them to sign for SSL sites within that domain, keeping the organizational identifier the same.
I'm really kind of surprised by the comments in this thread. It's almost like nobody remembers the fairly recent JPEG comment heap overflow problem in Netscape. I can't find the Slashdot comment right off-hand, but do a netsearch for more information. This issue is not that far-fetched, folks.
The original poster just went about it the wrong way in my opinion. It's like auto-insurance. Some people aren't happy with what they've got, but instead of trying to work it out, they first start shopping elsewhere and only "break the news" when they're ready to leave.
If you come straight out and say you think you deserve a little more, the company may not be as likely to respond favorably, but they also won't treat you like an employee that will be ready to bolt at the first sign of a marginally better prospect. Loyalty may be a lost thing to the big corporations, but it does make a difference to your immediate superiors, and generally they're the ones that have the say in whether you do get that raise or promotion.
But hell, if they turn you down, by all means shop around. Personally, I'm always on the look-out, but it's got to be significantly better than what I have before I'd entertain actually leaving.
Yes, I suspect it will. But hopefully some of that will be mitigated by careful application design. Make it easy or automatic for things to be kept secure and up to date (including best practice education). IIS and Windows are starting to try and move in that direction.
And I totally agree with you -- I'm not going to challenge the decision to go with IIS. In many cases there are perfectly reasonable business reasons to do that in some environments. I just think it's a trade-off.
Slashdot Mirror
To be fair, while I've never been in one of his tutorial sessions, I've found mjd to be a very bright guy and a very entertaining speaker. His presentations at YAPC were heavily attended, and I found them interesting and easy to follow, especially this one.
Would you rather he not have spoken up before the tutorial began, in this case? Let people throw their $800 away without knowing that they already know the material until it's far too late?
Regarding the lack of multi-window support, I agree. I've played with XFree86 for Win32 quite a lot, and this is the chief and only significant reason it's unsuitable for the things I do.
With everything in one big X window, I *have* to load an X window manager and have to keep all of my X apps contained in that one window. It makes it a little awkward if I need to be working between a Windows app and an X app, because I'm having to swap out of the X system entirely to go back to a Windows app. It's just a little cumbersome. I love the way Exceed is able to do this: every X app appears seamlessly like any other Windows app. And it's fast.
Plus, it's sometimes interesting to still use a window manager or shell in this mode. You get the benefits of having X windows float around a Win32 desktop, in addition to having these windows all managed by a different window manager. It's kind of cool having your Gnome panel, desktop icons, etc., floating on your Win32 desktop alongside your Explorer bar and Explorer desktop icons.
I disagree. If you're sticking something up on a web site, that something has a URL. Every entity on a web site has its own unique URL that should be retrievable anywhere.
If you don't like this behavior, and you want "pages" on your site to only be accessible by people browsing through your site, you're going to need to stick a "document retrieval" application layer onto your site. Users start a session when they enter this application, and are only able to retrieve stories through this application front-end. This can be done through HTTP as simply as with a session ID, but the web was not meant to work like this.
Again, we have a rather useful technology being twisted and warped by corporate interests instead of those corporate interests funding a proper technological solution, just like the intellectual property crap associated with DNS nowadays.
Other benefits of normalizing on a distance based on the speed of light is that c can become 1, greatly simplifying things like e=mc^2 to e=m^2, for instance. We can essentially just start eliminating the speed of light from equations entirely, since it will (generally) just factor out if it's equal to one.
We have a large number of units that depend directly upon the unit we use for time. Something fundamental as the Hertz would be redefined if we choose to use something new for a "metric" second (ignoring the fact that a "second" is already an SI unit).
The biggest problems with our method of measuring time (the requirement for leap years, leap seconds, etc.) won't be solved just by changing systems, since our planet doesn't rotate at a perfect rate and doesn't orbit the sun at an integral number of days.
You have to base a new system upon the rotation of the earth, because any time system that doesn't make sense relative to a single earth day would not be accepted. So while we can very nicely subdivide a day, a "day" is still pretty arbitrary and specific to the earth. When we deal with astronomical events, or have to tell time on another planet, we're still stuck with doing it like they do on earth.
It might make more sense from a scientific point of view to adopt a "second" that's more closely tied in "metric" respects to the half-life of a known element, or perhaps something like the absorption spectrum of hydrogen. (Normalizing to the latter, keeping the base unit as close as possible to a traditional second, a unit equivalent to exactly 1 billion times the absorption frequency of hydrogen would be about 0.7 traditional seconds, which would put closer to 123000 of them in a day.)
Units of distance should be measured (regardless of our definition of "second") in some form of light-seconds. Using our new and improved second, 1 nano-light-second would be equivalent to about 21cm (or nearly 8-1/2 inches). Not coincidentally, this also happens to be the wavelength of hydrogen absorption. Not too cumbersome, eh?
(Warning: I haven't checked my math too closely.)
Some states do have statutes prohibiting caller ID blocking for telemarketing calls. Unfortunately a lot of the telemarketing calls I receive are simply "OUT OF AREA" and are not blocked, just unavailable. Since I get one or two other regular people that are "OUT OF AREA" calling me, I have to pick these up as well.
Fortunately I'm on my state's no-call list, so those that do call me get an earfull.
Most (all?) laws regulating telemarketing have specific exemptions for companies you are doing or have recently done business with.
In short, it's unlikely you have any legal backing to keep your own bank from marketing their services to you on the number you've given them.
Same concept though.. There may not be the same type of demand then.
Agreed, even in my case where I have two monitors on a WinXP system, I have more than enough space to keep all of my background apps running *and* still be productive with Photoshop or some other big app.
However, I do also have a third cheap monitor off to the side dedicated to receiving syslog messages and other text bits of importance from my Linux systems. It might be nice to offload some of this onto an LCD screen that I can use a little more portably. It's not my screen's real-estate that I'm concerned about, it's my desk's!
Why totally impractical? Use 'ruptime' to gather the stats, and a stock LCD app to send the data to the display..
"good, common words domain names can generate hits on their own."
This is why we're in the mess we're in. Hostnames and URL's were meant for a time when the Internet was in the hands of the techies. When business and legal interests completely overwhelmed the purpose of the Internet, nothing better stepped in to supplement the current naming system.
We basically need a new directory service: one that maps legal entity names to DNS domains. Legislate this with all of the intellectual property crap you want, but leave DNS itself out of it. There's no reason users need to really see DNS domain names, and little reason they'd even need to see URL's for that matter. What percentage of the general public understands http:// and ftp://?
I should be able to tell my Internet browser to go to "IBM", and it should be able to consult a directory of entities, find a list of legal entities that have the name "IBM", obtain a DNS domain name from that directory (ibm.com), do an SRV lookup for the HTTP service at ibm.com, make the request for / at this server, and boom, you're at the "IBM" organization's home page. A TLS/SSL certificate at this point could validate the organization's identity (it should match the original directory entry), for those looking at security. Nowhere in this process do I have to guess that www.ibm.com is IBM's home page. The hostname in this case ceases to be such a valued commodity, outside of vanity uses.
We no longer see companies registering a hundred 2nd-level domains for every service mark and product they sell. For that type of thing, RealNames or similar services would (and do) work fine.
DNS names were never intended to be commodities (at least not to the degree they are today), any more than Java class name hierarchies or SNMP MIB organizational identifiers are meant to be commodities. DNS names have become commodities because there is presently no directory that associates legal business names to DNS domains. DNS is being warped into serving two purposes at once here: the association of names to IP addresses, and the association of legal business entities to an IP address for a web site.
What we need is to return DNS to serving its original role: to provide an easier-to-use addressing mechanism for Internet hosts. The role of associating legal entity names to Internet domains needs its own service (e.g. X.500 or LDAP). A "keyword" lookup service for product names or other service marks would need a third service a la RealNames.
We need to desperately curb the use of www.what-i-am-looking-for.com and to start enforcing DNS delegation like it was originally designed.
My two cents.
At first, before reading this into more detail, I was on the side of the Slashdot masses here. It seemed unreasonable that they'd be unwilling to share the list. What have they got to hide?
But then, after actually reading the decision, it made sense to me. By publishing the list, they would be basically providing to the world a nice, comprehensive list of, among other things, child porn sites, nicely categorized.
Now granted, child porn probably does not comprise the entirety of this list, but is it really worth turning the Australian government into the world's best child porn links site just to add your own level of oversight into their blocking?
I might suggest a couple of slightly different (non-mutually-exclusive) approaches to handling this trust issue:
1. Create a separate organization specifically to act as oversight for the list. Put them in a position to have access to this list, but keep the list itself out of general circulation. If you don't trust the government, maybe you can put some people you do trust into this oversight panel.
2. In addition to categorizing links to content as safe or unsafe, categorize the link itself along with it. If the group doesn't feel that link needs to be distributed, it can be excluded from general distribution (this solving the government's concerns over release). Everything else would make a publicly-available list and can enjoy all of the public scrutiny you like. This doesn't help if you believe the censors will be abusing their power by censoring things they shouldn't (since they'd just keep the link from you as well), but it would aid in catching accidental classifications, or where the classifications are likely to be contested.
Don't get me wrong, though, I don't really approve these acts of censorship either. But if the Australian government has decided to do it, they're going to do it. We can work to help them do it right without hindering the movement to keep them from doing it at all.
In addition, it should be fairly easy to identify censorship when you see it. To date I don't think I've read anything about how their censorship program has blocked something that shouldn't be. Maybe we can give them the benefit of the doubt (for now)?
Trademark law says the "own" it.
And this would be different from RealNames because RealNames would only allow the association of that one word to that one company, so it suffers some of the same problems as DNS.
A proper directory service wouldn't care. If your business's name is Yahoo, that would go into the directory in the proper place, and a search of the directory would give you what you're looking for. It's not the convenience of 1-word "keywords" in some cases, but it beats having to guess a hostname and, thus, giving hostnames a monetary value and a herd of companies and lawyers ready to exploit that.
I'm kind of curious about your line of reasoning here. I think DNS as it's designed does a perfectly adequate job at resolving names.
I think it's a good thing that only "A" records take you from a name to an IP address. All other record types should only permit a name. Administration would be annoying otherwise. Same with CNAME records. If I want to express the fact that www.company-name.com should be the same as www.company.com, and we're in different groups, we shouldn't have to coordinate IP address changes. I just set up a CNAME record telling everyone my "www" is really the other "www", and to go get the "A" mapping from there. (On a side note, I completely disapprove of using this method to alias one web server to the other, unless the destination web server issues a redirect to the canonical web site URL.)
DNS is for providing information about domain names, including (but not limited to) IP addresses. You also have service information (e.g. MX and SRV records), contact information, etc. Any piece of data you need to associate with a DNS domain name should be expressible with the DNS protocol.
DNS as it is does its job fine, I have no problem with that. It's the fact that we're trying to squeeze intellectual property status to these assigned names that's the problem.
It's no more immense than DNS (which, admittedly, is pretty freakin' immense). X.500 (and LDAP) is designed to be distributed. You'd have a root (like DNS), but no "generic" top-level domains. You'd just hang each country's X.500 directory off of the root, and each country would be responsible for getting its local organizations listed in its directory. For the US, this could mean sub-delegations at the state level, sort of like how the .us DNS domain works. Traffic between tiers should be authenticated, a la DNSSEC (but who uses that?). When a business (or individual, for that matter) gets registered, they go into their local X.500 (or LDAP) directory and are immediately visible to the globe. This has the added benefit of doubling as a global telephone or address directory, with whatever information the user wishes to publish.
But really I'm kind of digressing. This is just one potential method for accomplishing the same goal: reducing DNS back to what it was designed for and creating something else to fill the demand for a look-up service, designed for that purpose.
Somebody else mentioned RealNames, and that's a step in the right direction, but it's still trying to use the "one word" approach to resolving trademarks. Why should 'Apple Computer' have any more claim over 'Apple' than 'Apple Hardware Store'?
Expanding upon this comment, the directory service should not even be web-centric. It should just associate a legal business name with the assigned DNS domain name (and/or telephone number, physical address, whatever). Using DNS SRV records, you could then associate a web server or an internal directory server with this domain name. All of this can be done without having to guess that "www.company-name.com" is the home page for a company. You'd just ask for "Joe's Widgets", and you'd get back joes-widgets.co.yz.us (or whatever). This domain would have an SRV record pointing to perhaps web.it.joes-widgets.co.yz.us, which would be a web server located in Joe's Widget's IT group. A certificate presented by the server would authenticate it as Joe's Widgets of Anytown, YZ, as certified by the state of YZ, as certified by the US, and your location effort would be complete.
I mentioned search engines.
But search engines as they are today have one flaw: it's difficult to find authoritative information.
If you want to find organizational information (e.g. a web site) for, say, Joe's Widgets located in Anytown, YZ, you might do a netsearch like you imply. This netsearch might come back with one or more web pages related to this company, and if you're lucky, the first link might even be the home page for this company itself.
But realistically, we don't have a "phone book" equivalent for the Internet. That's what a real directory service can provide that search engines, today, do not.
I'm not saying that commercialization of the Internet should not have occurred, I just think that steps should have been taken early on to allow that to happen in a more sane fashion. Treating the DNS hierarchy as an intellectual property battle-ground is silly. DNS domains should have exactly the same weight as any assigned name or number.
What are we going to see next? IP battles for control over different Java class name hierarchies? What about SNMP MIB organizational identifiers?
DNS is a technological tool, and should not be a business tool (like it's become).
I'm interested in reading the book's take on this, but it seems to me that DNS was never meant to be twisted into what it's being used for today.
DNS's primary purpose was to allow someone to lookup a numerical address from a hierarchial alphanumeric one. It was never meant to be a search mechanism for people to find an organization. That's what directories like X.500 and LDAP, or search engines are for. The very fact that there can be multiple commercial entities primarily identified with the same word makes DNS wholly unsuitable for this.
I think we need to scrap DNS and start over from scratch. Hand out names as needed for technological reasons, and set up a directory for users to provide an organizational name and get back links to that organization's web site(s). Regulate that directory with all of the intellectual property and trademark crap that you want, but do it in such a way that local IP and trademark law can be respected. Hide DNS (and maybe URL's entirely for that matter) from the user through their browsers, and utilize certificates for establishing identity, instead of the basic assumption that if the domain name in the URL "looks right", then it must be them. I might even go so far as to suggest that when an organizational's DNS domain is created, they automatically get a certificate that allows them to sign for SSL sites within that domain, keeping the organizational identifier the same.
I'm really kind of surprised by the comments in this thread. It's almost like nobody remembers the fairly recent JPEG comment heap overflow problem in Netscape. I can't find the Slashdot comment right off-hand, but do a netsearch for more information. This issue is not that far-fetched, folks.
http://online.securityfocus.com/bid/1503
The original poster just went about it the wrong way in my opinion. It's like auto-insurance. Some people aren't happy with what they've got, but instead of trying to work it out, they first start shopping elsewhere and only "break the news" when they're ready to leave.
If you come straight out and say you think you deserve a little more, the company may not be as likely to respond favorably, but they also won't treat you like an employee that will be ready to bolt at the first sign of a marginally better prospect. Loyalty may be a lost thing to the big corporations, but it does make a difference to your immediate superiors, and generally they're the ones that have the say in whether you do get that raise or promotion.
But hell, if they turn you down, by all means shop around. Personally, I'm always on the look-out, but it's got to be significantly better than what I have before I'd entertain actually leaving.
Yes, I suspect it will. But hopefully some of that will be mitigated by careful application design. Make it easy or automatic for things to be kept secure and up to date (including best practice education). IIS and Windows are starting to try and move in that direction.
And I totally agree with you -- I'm not going to challenge the decision to go with IIS. In many cases there are perfectly reasonable business reasons to do that in some environments. I just think it's a trade-off.
Einstein also did fairly poorly in math, as I recall.
Einstein was a brilliant theoretical physicist, but in other areas he was just a man. Please don't take all of his quotes as gospel.
Two minutes is a long time when another person can do a similar calculation in his head in a few seconds.