as far as I can see Eazel's stuff is a pure service, not an after-sale, since there is NO SALE to begin with. and since their software is open source, they sure can't be accused of tying a product (whatever they develop) with a service. hence, for all I care, they should feel more than free to base their business on offering their update services -- although I have my reservations as to whether it's a viable model. and if it *does* work, it'll be precisely because of this lack of tying; once the users know that they're free to use whatever service (or none), and that this is freedom is protected by the fact that the software is open source, they may actually have less reservations to use the service in question.
well if that is true, then it's *incredibly good news*, since it would mean that all restrictive software EULA-ing would become invalid (find itself without any legal foundation), except of course for those licenses that cover copying only (so the GPL is actually safe).
To me, all of this seems like trying to rip off companies that are
providing something which people obviously want. And if people
succeed, then these companies are going to suffer, which means no more
deals for people. Is this what we want?
yes. this is what we (or at least, I) want. I want after-sale business models to FAIL so that companies can get back to the good old way of selling stuff without tying the user to some badly thought-out service of theirs. namely, by selling appliances at a profit. just like VCRs, stereos, cars, and everything out there that gets sold. selling hardware at a loss and recouping on side channels is BAD FOR THE CONSUMER.
built in firmware doesn't come with "licensing agreements" by any stretch of the imagination, for the very simple reason that there is no copying involved. the (really bad) excuse that opened the software licence floodgates was that software gets copied from disk to memory at the user's request, so the user needs a 'license' for that. no such thing happens with firmware.
A company trying to make money! The nerve! Seriously, of Mr. Blount's goals, how many are realistic for a company trying to make a profit? 1) Not paying for the service. Buy the box, nothing more.
how they make money is their problem, not mine. I don't pay money to use a VCR once I've bought it, and the TiVo is not much more than a glorified digital VCR. Sell it for whatever it costs to make it plus a reasonable margin, and be done with it! It's worked for all products up to now, why does suddenly everyone want users to *subscribe* to things, and why would we users buy into that?
2) Privacy. TiVo's privacy policy is pretty well defined on their web page ( http://www.tivo.com/care/privacy.html ). Yes, they sell info about what TiVo users are watching, but only at the zip code level. I'm comfortable with their privacy policy.
are you comfortable with them changing it at any time, like Amazon just did?
4) Skip all ads. ReplayTV took the approach of "screw the networks, we don't need 'em" and added a 30 second skip button. Since then, they've hired a former CBS exec and have realized that they will have to work with the networks in the future. The skip button is still there, but it is never advertised.
hey, at least it *is* there; I couldn't care less it's advertised or not, but if I'm going to buy a digital vcr (which i'm not, but that's another story), I sure as shit want the one with the "skip ad" button.
I don't understand all of the criticism towards TiVo. Yes there are features we would like to see but this is still a fairly new market. For a product that has only been available for less than two years or so, I think it's a very impressive device. Plus, you can upgrade 'em to 100 hours plus with relative ease.:)
the original article makes exactly this point. TiVo is nice, and a huge step up from what was available before. but it's not perfect, and not only in features but also in the model of interaction between the user and the manufacturer. I don't think it's reasonable to criticize TiVo much; they've done a good hack and sold it. If we don't like the conditions, it's up to us to work towards being able to run our own personal TiVo-like services from a regular computer using open source software. It may be a year or maybe three, but I'm pretty sure it'll come.
hey, cut them some slack... a minimum dose of marketese is pretty much required on press releases. they're doing great stuff, and their actual site looks very usable; that's what counts.
How do you distinguish between "the world itself" and our thoughts (specifically, our perceived impression of the world.) Our only knowledge of the world is that shown to us by our model of the world.
That's a question that ultimately can't be answered, because any answer is going to be made of sentences which relate to thoughts and models, not to the "world itself". The whole
idea of "the world itself", as it occurs in our heads is a shared, largely unconscious model which describes such basic stuff as "it hurts if I walk into you". You can't ask me to think outside of my head, but that doesn't mean I'm going to stop believing in the physical world anytime soon. And I assumed this belief in writing my comment.
Anyway, here's an example of something that is (in my understanding) not a model or thought, nor mediated by the senses: a headache. When I have a headache, I don't need a model of any kind to feel it. Of course, I *do* have a mental model of what a headache is like, and how paracetamol helps alleviate one, but if I didn't have the model, I'd still feel the headache, I'd just be confused about it. The lowly headache is an example of pure direct perception. Someone stop me before I start a cult to the almighty headache;)
ironically enough, wanting our thoughts to be free of any culture's influence is very much a western-modern thing:)
me, I'd rather shoot for another way of opening my mind: trying to *really* understand at least two radically different cultures, from the inside (yeah, i'm working on it, and it's quite confusing at times). of course, that is not culture-free either in itself; it's another western-modern trend to value multi-culturalism, just a somewhat less mainstream one:=)
I realize that you are joking, but what you say is not all that far from how newborns start making sense of their senses. it is generally believed that they don't see "objects" from the start, but blobs of light and color (and this is also what adults see when they've been blind for all their life and somehow have their vision restored -- reference: one of Oliver Sachs' books). if you think of it, the whole idea that the world is made of "things" is a mental construct -- the actual, underlying physical world (assuming materialism here) only has particles and fields and the like, there is nothing in an electron that makes it part of "this table". of course, our senses don't deal with the world at the particle level, but the same problem applies: your brain has to *learn* that the table is one "thing" and the book on it is another "thing" even though they're physically contiguous.
yep. the whole success of the Whorfian hypothesis may however be indirect proof of its own validity, *considerably weakened*. consider the fact that the Whorfian idea survives very much in popular culture, even though it has pretty much been proved (as you say) to be false. couldn't that be related, at least in part, to the fact that it has a name?
hmmm.. yes and no. I completely agree that models and metaphors ARE what our *thoughts* are made of (that is, our discursive thoughts, and quite a good part of the rest -- 'intuition' and the like).
however, I think you are largely overstating the supposed doubtlessness of your stronger proposition, that models and metaphors are the world itself. it makes sense in a po-mo context, but if you think academia's opinion-of-the-week on the nature of reality is 1) universally accepted across schools of thought and cultures, and 2) not going to change in a decade or two, then well... i'll have to respectfully disagree with you.
more generally, I'd say it's always good to weigh good old common sense in front of the latest theory. postmodernism has revealed some rather large gaps in so-called objectivity, and I'm the first to rejoice of it -- but I do'nt necessarily buy its stronger claims, and I do think that the word "truth" still *sometimes* makes sense.
it's very common to have bad experiences with code outsourcing. i know a couple of examples first hand, where a company hired another to write code for a relatively large subsystem, and the thing was delayed again and again for months, and then finally the coders either came up with something so broken that it couldn't be used, or just threw the towel. the original company ended up implementing the things themselves, in a much simpler way that actually worked.
OTOH, I'm sure that if you shop really well for an outsourcing place, work with them on specs and requirements, and agree on a reasonable delivery time, you have a much better change to get good results.
one interesting possibility would be to use open-source-like development methods. the code doesn't need to be open source, or available to the public, but you can set up development in the typical OSS way, with all communication between developers being done on mailing lists, with an internal CVS tree, and encouraging everyone to keep an eye on the patches, and to read other people's code. then you can have one or two people from the "client" company on the project, while the "outsourced" company throws as many developers at it as it considers appropriate. since there is constant contact at the techie level between the two companies, you greatly diminish the possibilities of "bad surprises".
amazon doens't have your credit card number of type or name while you're filling your shopping cart, only afterwards when you sign in your info or use a previously entered profile. or rather, the only way amazon can have a clue who you are is with cookies from previous sessions. so rename your cookie file and try again; if you get a better deal, great; if you get a worse one, you can always put your cookies back.
there may not be any FSF code in there (I haven't checked), but there sure is non-KDE-specific code. there's even some GPL code of my own in one of the KDE modules, taken from an old project of mine that had nothing to do with KDE. not that I'm complaining or anything, I found it nice to see my code in use... but I do think that they're twisting things a bit when they say that only "a few bits" were written elsewhere.
anyway, it's pretty sad that even now that Qt is under the GPL, people (read: RMS) feel the need to be divisive about licensing issues still. many companies and teams have had licensing troubles (remember Corel's beta program?) and then cleaned up their act. KDE is now 100% clean, let's forget the QPL fiasco and enjoy the increased choice and competition between KDE, GNOME, XFCE and others.
This is one of those
things that you (and I and everyone else) just has to accept. You and
I know that email and usenet support would be just fine, but
management will never believe it. Management needs to pay for things
to feel secure about it, and above all, they need to be able to point
their fingers at something.
you're plainly admitting here that your management is just being obtuse, over-conservative, and that you *know* it would work. in other words, it's your management's problem. not linux's problem, and certainly not the problem of every other management on earth. and btw I can point you to more than a few production linux or bsd-based boxes that don't have a support contract from anyone. guess who's saving money and making smart decisions, and who isn't.
I have a feeling that many people are like me, in that they don't
always want to be found.. I don't want my work calling me at 10pm,
after i've worked 10 hours, and am out on a relaxing walk with my
girlfriend..
yup, my feelings exactly. that's the reason why I didnt' want a cellphone in a very long time, and then when they got cheap enough I finally got one, which is *always* turned off, just for those rare cases when I go somewhere and want to be able to call. no-one has my cell number; for that matter, I don't remmeber it myself. and I don't pay a subscription, it's one of these card GSM phones where you put money into the card, and pay nothing if you do'nt use it. it's not cost effective if you use it to call a lot, but i don't, so that's fine with me.
that goes well with that one line in the article: ""The importance of brand becomes that much more of a factor".
anyway, I can't say I'm surprised about sales dropping down. after all, once you have a cellphone, you don't need another unless the first breaks *or* you're changing for 'image' reasons. and when the best reason to change is 'image', as far as I'm concerned, it means that the field has become yet another plain old commodity. i may want to have one or not, but what's clear is that once i do have one i'm not going to buy another for a very long time.
i mean, sure,/bin/ash is not the best known application out there, but I don't think it's *that* obscure either.... or does it have some dark origins than 'they' don't want us to know of?
NAT is conceptually ugly because it breaks IP's basic rule "only the endpoints should know or care about established connections". but hey, it works, and it works pretty well. Of course, NAT alone doesn't make a firewall... but it sure gives you a convenient place to put your firewalling rules, and *then* it can be one.
reverse proxy can be done with Apache and mod_proxy, see the documentation for mod_proxy at http://www.apache.org/docs/mod/mod_pr oxy.html. To do name-based vhosting with it, you have two options: either have the <Virtualhost> directives on the rev-proxy and forward to different URL paths on the backend (i.e www.bletch.com/urf becomes backend.serverfarm.com/bletch/urf), or you pass the Host: field as Original-Host to the backend, and then setup a fixuphandler to put it back as the Host: header. There is an example module that does something similar (passes the original request's IP as X-Forwarded-For" at
http://www.cpan.org/auth ors/id/ABH/mod_proxy_add_forward.c; it's originally meant for use with mod_perl, but there's no reason why it wouldnt work with anything else on the backend, with a tiny bit of C hacking.
that seems the most likely. the only other place I've ever seen such botched http (including the "reply after the first line" crap) was from a java program that took http requests directly (but whose main job wasn't serving webpages). I'm sure the IIS guys have read the http rfc's a little more carefully than this. anyone care to run one of those tcp signature OS-guessing programs on www.hotmail.com?
are you deliberately lying, or does your ISP put you behind a transparent proxy without telling you? when I send a "HEAD / HTTP/1.0" to slashdot.org, with a single CRLF, it just sits there waiting, and times out (closes the connection w/o reply) after 25 seconds or so. I sure don't see it responding after the first CRLF pair.
"Simple-Request and Simple-Response do not allow the use of any header
information and are limited to a single request method (GET).
Simple-Request is HTTP 0.9. A modern webserver has no business doing HTTP 0.9 replies to HTTP 1.0 requests, and anyway www.hotmail.com is not claiming 0.9 but 1.1 in its response. So the fact that http 0.9 doesn't do http headers is irrelevant here.
HTTP/1.0 does NOT require the server to wait for two CRLF's before
sending a response. Try the same thing on Slashdot -- or try reading
the RFC1945 documentation on HTTP/1.0.
Once again, www.hotmail.com's IIS is claiming http 1.1 in its response, so it's rfc2616 (not 1945 or 2068) that applies there. Anyway, I'll concede this point -- I'm not aware of a strict requirement to read the entire request before answering, and I can't be bothered to skim rfc2616 for one (though I wouldn't be surprised if such a requirement was an indirect consequence of some compulsory headers). OTOH, slashdot.org does wait for a \r\n\r\n.
Also, HTTP/1.0 does NOT require the server to send two CRLFs at the end of a transmission: the format is: Full-Response = Status-Line *( General-Header ; Section 4.3 | Response-Header ; Section 6.2 | Entity-Header ) ; Section 7.1 CRLF [ Entity-Body ] ; Section 7.2
And if you look a but further down on rfc1945, in section 4.2 you see that the *-Header elements end in a CRLF. together with the standalone CRLF in Full-Response, that means two successive CRLFs, before the (optional) Entity-Body.
So it does look like http 1.0 (and 1.1 too) requires \r\n\r\n at the end of the headers.
And by the way, Connection: close is only required for HTTP/1.1 connections
and www.hotmail.com is botching up these 1.1 connections: $ socket -cv www.hotmail.com 80 connected to www.hotmail.com port 80 (www) HEAD / HTTP/1.1 HTTP/1.1 302 Redirected Server: Microsoft-IIS/5.0 Date: Tue, 29 Aug 2000 19:39:59 GMT Location: http://lc2.law5.hotmail.passport.com/cgi-bin/login
this is wrong on many counts. first, my request was invalid since it didn't include a Host: header, yet www.hotmail.com accepted it (and woulnd't have read my Host: even if I'd sent one, since it answered after the first line). Second, by rfc2616 seciton 14.10, a server that doesn't support persistent connectuls MUST send a "Connection: close" line. Third, the missing \r\n at the end. Fourth, even if I use GET rather than HEAD, I still don't get an entity body or a content-type, which is not strictly required for redirects, but it's sloppy not to have one.
what surprises me about all of this is WHY this server sends so much crap. I can't believe IIS 5.0 by default botches the protocols so much, and other IIS servers I've tried seem to not have most of these obvious problems. I don't know what kind of bad hack the www.hotmail.com guys did, and frankly, I'd rather not know.
Slashdot Mirror
as far as I can see Eazel's stuff is a pure service, not an after-sale, since there is NO SALE to begin with. and since their software is open source, they sure can't be accused of tying a product (whatever they develop) with a service. hence, for all I care, they should feel more than free to base their business on offering their update services -- although I have my reservations as to whether it's a viable model. and if it *does* work, it'll be precisely because of this lack of tying; once the users know that they're free to use whatever service (or none), and that this is freedom is protected by the fact that the software is open source, they may actually have less reservations to use the service in question.
well if that is true, then it's *incredibly good news*, since it would mean that all restrictive software EULA-ing would become invalid (find itself without any legal foundation), except of course for those licenses that cover copying only (so the GPL is actually safe).
built in firmware doesn't come with "licensing agreements" by any stretch of the imagination, for the very simple reason that there is no copying involved. the (really bad) excuse that opened the software licence floodgates was that software gets copied from disk to memory at the user's request, so the user needs a 'license' for that. no such thing happens with firmware.
how they make money is their problem, not mine. I don't pay money to use a VCR once I've bought it, and the TiVo is not much more than a glorified digital VCR. Sell it for whatever it costs to make it plus a reasonable margin, and be done with it! It's worked for all products up to now, why does suddenly everyone want users to *subscribe* to things, and why would we users buy into that?
are you comfortable with them changing it at any time, like Amazon just did?
hey, at least it *is* there; I couldn't care less it's advertised or not, but if I'm going to buy a digital vcr (which i'm not, but that's another story), I sure as shit want the one with the "skip ad" button.
the original article makes exactly this point. TiVo is nice, and a huge step up from what was available before. but it's not perfect, and not only in features but also in the model of interaction between the user and the manufacturer. I don't think it's reasonable to criticize TiVo much; they've done a good hack and sold it. If we don't like the conditions, it's up to us to work towards being able to run our own personal TiVo-like services from a regular computer using open source software. It may be a year or maybe three, but I'm pretty sure it'll come.
hey, cut them some slack... a minimum dose of marketese is pretty much required on press releases. they're doing great stuff, and their actual site looks very usable; that's what counts.
That's a question that ultimately can't be answered, because any answer is going to be made of sentences which relate to thoughts and models, not to the "world itself". The whole idea of "the world itself", as it occurs in our heads is a shared, largely unconscious model which describes such basic stuff as "it hurts if I walk into you". You can't ask me to think outside of my head, but that doesn't mean I'm going to stop believing in the physical world anytime soon. And I assumed this belief in writing my comment.
Anyway, here's an example of something that is (in my understanding) not a model or thought, nor mediated by the senses: a headache. When I have a headache, I don't need a model of any kind to feel it. Of course, I *do* have a mental model of what a headache is like, and how paracetamol helps alleviate one, but if I didn't have the model, I'd still feel the headache, I'd just be confused about it. The lowly headache is an example of pure direct perception. Someone stop me before I start a cult to the almighty headache ;)
why don't you just run a ssh daemon on your home machine, on a port that they do allow, like the telnet or http one?
me, I'd rather shoot for another way of opening my mind: trying to *really* understand at least two radically different cultures, from the inside (yeah, i'm working on it, and it's quite confusing at times). of course, that is not culture-free either in itself; it's another western-modern trend to value multi-culturalism, just a somewhat less mainstream one :=)
I realize that you are joking, but what you say is not all that far from how newborns start making sense of their senses. it is generally believed that they don't see "objects" from the start, but blobs of light and color (and this is also what adults see when they've been blind for all their life and somehow have their vision restored -- reference: one of Oliver Sachs' books). if you think of it, the whole idea that the world is made of "things" is a mental construct -- the actual, underlying physical world (assuming materialism here) only has particles and fields and the like, there is nothing in an electron that makes it part of "this table". of course, our senses don't deal with the world at the particle level, but the same problem applies: your brain has to *learn* that the table is one "thing" and the book on it is another "thing" even though they're physically contiguous.
yep. the whole success of the Whorfian hypothesis may however be indirect proof of its own validity, *considerably weakened*. consider the fact that the Whorfian idea survives very much in popular culture, even though it has pretty much been proved (as you say) to be false. couldn't that be related, at least in part, to the fact that it has a name?
however, I think you are largely overstating the supposed doubtlessness of your stronger proposition, that models and metaphors are the world itself. it makes sense in a po-mo context, but if you think academia's opinion-of-the-week on the nature of reality is 1) universally accepted across schools of thought and cultures, and 2) not going to change in a decade or two, then well... i'll have to respectfully disagree with you.
more generally, I'd say it's always good to weigh good old common sense in front of the latest theory. postmodernism has revealed some rather large gaps in so-called objectivity, and I'm the first to rejoice of it -- but I do'nt necessarily buy its stronger claims, and I do think that the word "truth" still *sometimes* makes sense.
OTOH, I'm sure that if you shop really well for an outsourcing place, work with them on specs and requirements, and agree on a reasonable delivery time, you have a much better change to get good results.
one interesting possibility would be to use open-source-like development methods. the code doesn't need to be open source, or available to the public, but you can set up development in the typical OSS way, with all communication between developers being done on mailing lists, with an internal CVS tree, and encouraging everyone to keep an eye on the patches, and to read other people's code. then you can have one or two people from the "client" company on the project, while the "outsourced" company throws as many developers at it as it considers appropriate. since there is constant contact at the techie level between the two companies, you greatly diminish the possibilities of "bad surprises".
amazon doens't have your credit card number of type or name while you're filling your shopping cart, only afterwards when you sign in your info or use a previously entered profile. or rather, the only way amazon can have a clue who you are is with cookies from previous sessions. so rename your cookie file and try again; if you get a better deal, great; if you get a worse one, you can always put your cookies back.
anyway, it's pretty sad that even now that Qt is under the GPL, people (read: RMS) feel the need to be divisive about licensing issues still. many companies and teams have had licensing troubles (remember Corel's beta program?) and then cleaned up their act. KDE is now 100% clean, let's forget the QPL fiasco and enjoy the increased choice and competition between KDE, GNOME, XFCE and others.
spanish and catalan, too.
you're plainly admitting here that your management is just being obtuse, over-conservative, and that you *know* it would work. in other words, it's your management's problem. not linux's problem, and certainly not the problem of every other management on earth. and btw I can point you to more than a few production linux or bsd-based boxes that don't have a support contract from anyone. guess who's saving money and making smart decisions, and who isn't.
anyway, I can't say I'm surprised about sales dropping down. after all, once you have a cellphone, you don't need another unless the first breaks *or* you're changing for 'image' reasons. and when the best reason to change is 'image', as far as I'm concerned, it means that the field has become yet another plain old commodity. i may want to have one or not, but what's clear is that once i do have one i'm not going to buy another for a very long time.
i mean, sure, /bin/ash is not the best known application out there, but I don't think it's *that* obscure either.... or does it have some dark origins than 'they' don't want us to know of?
NAT is conceptually ugly because it breaks IP's basic rule "only the endpoints should know or care about established connections". but hey, it works, and it works pretty well. Of course, NAT alone doesn't make a firewall... but it sure gives you a convenient place to put your firewalling rules, and *then* it can be one.
reverse proxy can be done with Apache and mod_proxy, see the documentation for mod_proxy at http://www.apache.org/docs/mod/mod_pr oxy.html. To do name-based vhosting with it, you have two options: either have the <Virtualhost> directives on the rev-proxy and forward to different URL paths on the backend (i.e www.bletch.com/urf becomes backend.serverfarm.com/bletch/urf), or you pass the Host: field as Original-Host to the backend, and then setup a fixuphandler to put it back as the Host: header. There is an example module that does something similar (passes the original request's IP as X-Forwarded-For" at http://www.cpan.org/auth ors/id/ABH/mod_proxy_add_forward.c; it's originally meant for use with mod_perl, but there's no reason why it wouldnt work with anything else on the backend, with a tiny bit of C hacking.
that seems the most likely. the only other place I've ever seen such botched http (including the "reply after the first line" crap) was from a java program that took http requests directly (but whose main job wasn't serving webpages). I'm sure the IIS guys have read the http rfc's a little more carefully than this. anyone care to run one of those tcp signature OS-guessing programs on www.hotmail.com?
are you deliberately lying, or does your ISP put you behind a transparent proxy without telling you? when I send a "HEAD / HTTP/1.0" to slashdot.org, with a single CRLF, it just sits there waiting, and times out (closes the connection w/o reply) after 25 seconds or so. I sure don't see it responding after the first CRLF pair.
$ socket -cv www.hotmail.com 80
connected to www.hotmail.com port 80 (www)
HEAD / HTTP/1.1
HTTP/1.1 302 Redirected
Server: Microsoft-IIS/5.0
Date: Tue, 29 Aug 2000 19:39:59 GMT
Location: http://lc2.law5.hotmail.passport.com/cgi-bin/logi
this is wrong on many counts. first, my request was invalid since it didn't include a Host: header, yet www.hotmail.com accepted it (and woulnd't have read my Host: even if I'd sent one, since it answered after the first line). Second, by rfc2616 seciton 14.10, a server that doesn't support persistent connectuls MUST send a "Connection: close" line. Third, the missing \r\n at the end. Fourth, even if I use GET rather than HEAD, I still don't get an entity body or a content-type, which is not strictly required for redirects, but it's sloppy not to have one.
what surprises me about all of this is WHY this server sends so much crap. I can't believe IIS 5.0 by default botches the protocols so much, and other IIS servers I've tried seem to not have most of these obvious problems. I don't know what kind of bad hack the www.hotmail.com guys did, and frankly, I'd rather not know.