weren't there two competing 56K modem systems before V90 with V90 being the attempt to try and find a standard that both camps could upgrade thier firmware to?
using try-catch to detect null pointer dereferences isn't really safe though, what if the code is used on a CPU without a MMU?
maybe if you only target one CPU you can get away with this but its certainly very bad programming practice.
ultimately automated flaw scanners are like compiler warnings, they WILL create false positives but they should at the very least be carefully considered.
the only reasonable thing to do is to workarround even if its not your bug, because forcing everyone to upgrade the lib is often unfeasible.
sadly many developers really do not wan't to pollute thier code with workarounds for third party problems even if there is no other reasonable way to fix the bug for thier users.
The problem is that a good, solid DNS server these days won't answer recursive lookups to the world since that can be a way to game a dns server, or just overwhelm it into submission. lots of the alternate root systems run open public recursive resolvers, as long as the availibility of a few extra tld's doesn't bother you theese can be a good choice.
This is all because AOL dropped dialup service. (Could you ever get it in the UK? There must have been an equivalent.) AOL has always offered dialup in the uk and still does but unlike most providers they didn't (at least under the AOL name) go into the subscription free make the money from the phone calls buisness.
Why? It will turn out to be a good deal for them - someone else does all the upfront leg work and spending to put out a product - if they develop a hit, the big media companies can come in and sign them. the bad news is if someone builds up sufficiant fanbase the record companies end up fighting over them. This means that unless there is collusion involved the band will get a far better deal than one coming from nothing and this will cut into the record companies profits.
afaict the bands that make the record companies the biggest profits are the big hits that are still locked into thier initial (signed when they were unknown) contracts.
so china gets its oil slightly cheaper and the west slightly more expensive big deal
oil is a commodity, an expensive one but still a commodity. As such a single supplier cant really threaten a single customer (they can stop exporting thier oil altogether but that would hurt all oil customers as well as thier own pockets)
The facts of the world here in fact have not changed. Only nomenclature has. well what has actually happened is that nomenclature has caught up with the facts. The new(ish) fact is that there is a big belt of pluto like objects in similar orbits to pluto. The final straw came when it was discovered that pluto wasn't even the biggest in that belt.
It's also useful in education, because we ask our kids to learn the names of the planets, not every body that orbits the sun. There is really very little useful value in writing new textbooks here. well you could say there is very little value in teaching our kids the makeup of the solar system at all. After all its not as though any significant number of people leave earth and the only bodies with significant impact on everyday life are the sun and moon.
what we have really discovered here is that pluto was not a one of a kind in a pretty unique orbit but part of a belt of very similar lumps of rock. School textbooks talk about the asteroid belt but not ceres in particular. Similarly they should talk about the kuiper belt but not pluto in particular.
Personally I reckon they should auction names rather than selling them at a flat rate. for initial sales maybe and maybe renewals could be locked to the auction price but there is no way there should be an auction at renewal time.
domain name changes are very painfull for all involved and allowing anyone with the money to force someone else to undergo one on a whim is simply unacceptable and would tear apart the largely hobbyist structure of the net.
bad idea, one important design idea of domain names is they are supposed to be permanent (or as close to permanent as is reasonable) identifiers.
allowing a rich player to extort away a domain name from its long time user by outbidding them on registration fees would bring cahos to the net.
lets not replace an annoyance (cybersquatting) with something that will totally destroy the integrity of the naming system by allowing anyone to steal anyone elses name by paying a registration body more than the names owner can afford.
the government has a huge ability to both bribe (lots of money) and threaten (the government can very easilly make your life a misery for many years or of course they could just kill you and claim it was an accident).
punishing the companies is pointless for a descision that didn't involve the shareholders, it just hurts customers and shareholders alike while not really touching the execs. maybe the execs involved should be punished but i don't think even that is very fair given the pressure a government can introduce on an individual (any bribes should certainly be recovered though).
btw is the intel C compiler issue with amd chips (the one where non intel chips are sent down a slow codepath despite having the required feature bits to say they can use the fast one) still ongoing?
i should probablly clarify by accident before someone else does.
iirc it was discovered because an error in the then current understanding of orbital mechanics caused astronomers to belive there was a large object beyond neptune and pluto was discovered approximately where that large object would have been.
Heretofore a planet was (loosely) defined as a large mass in orbit around a star the problem with that definition is that "large" is not well defined.
since the switch to the heliocentric model a planet has generally meant a large relatively unique (within the solar system at least) object that orbits the sun. Other objects orbiting the sun were given other names (e.g. asteriods, comets, kuiper belt objects)
the problem is like ceres we are discovering pluto is not a planet in that sense but part of a large belt of similar objects that while significant as a group don't really have anything to set them apart from each other. Like ceres it should probally be reclassified.
the reason its such an emotive issue is that pluto was discovered by accident long before the belt in which it lies was discovered. so it has been in grade school textbooks as a "planet" for a long time.
i know its a diffiult balance, if you treat lusers like lusers then they will never switch unless forced, if you don't treat luesrs like lusers they get thier systems full of malware. Its a loose/loose choice.
what cds you wan't depends on the number of machines your installing and your ease of access to the internet during installations.
if you are planning to work away from an internet connection get the whole damn set of main CDs/DVDs.
if the machines don't have CD drives get the boot root and net-drivers floppies
if you are just doing one box and don't plan to use the cds after initial install get the netinst CD
if you wan't it to fit on a buisnesscard and don't care about ending up with an 486 optimised kernel (you can install an optimised one later from the net if you wan't) get buisnesscard.
installing stuff provided by your distro is easy, arguablly easier than installing stuff on windows (you don't have to go searching for download sources etc first)
the problem comes when you wan't something more custom. Something from outside the distros "walled garden" the distros are very resistant to universal packaging efforts like autopackage.
But since we are not using infinite number of decimal places, you are basically asking for a bigger float no? NO
what matters in money is predictable behaviour according to specified rules, those rules are decimal based so the system needs to operate with a (generally fixed) decimal exponent.
bigger binary exponent floats will give a better approximation of 0.01 but the error is still there and will still add up over time.
I am positive that you will find more failures in people that are (a) very mobile and (b) have only one adapter. yeah there are three things that stress connections afaict
1: use at full stretch (so any movement of anything gives heavy stress) 2: rough extraction from a tangle (so keep your cables untangled if possible, and if not possible at least try to be gentle with thier removal) 3: packing/unpacking, cables nearly always leave in directions that mean that they will have to either bend sharply to allow for use or bend sharply to allow the thing to be packed up reasonablly small.
which works fine until someone actually takes you to court and the court discovers that you used that reasoning and did a customer severe harm in the process.
thats what huge damages awards are really about, punishing corparations who use this line of reasoning.
weren't there two competing 56K modem systems before V90 with V90 being the attempt to try and find a standard that both camps could upgrade thier firmware to?
using try-catch to detect null pointer dereferences isn't really safe though, what if the code is used on a CPU without a MMU?
maybe if you only target one CPU you can get away with this but its certainly very bad programming practice.
ultimately automated flaw scanners are like compiler warnings, they WILL create false positives but they should at the very least be carefully considered.
what if "package B" is a third party lib though?
the only reasonable thing to do is to workarround even if its not your bug, because forcing everyone to upgrade the lib is often unfeasible.
sadly many developers really do not wan't to pollute thier code with workarounds for third party problems even if there is no other reasonable way to fix the bug for thier users.
The problem is that a good, solid DNS server these days won't answer recursive lookups to the world since that can be a way to game a dns server, or just overwhelm it into submission.
lots of the alternate root systems run open public recursive resolvers, as long as the availibility of a few extra tld's doesn't bother you theese can be a good choice.
This is all because AOL dropped dialup service. (Could you ever get it in the UK? There must have been an equivalent.)
? promo=228937&promoCode=228937
AOL has always offered dialup in the uk and still does but unlike most providers they didn't (at least under the AOL name) go into the subscription free make the money from the phone calls buisness.
they still offer an unlimited dialup (no phone call charges) package here in the uk but its more expensive than thier basic broadband package. http://info.aol.co.uk/dial-up/anytime-dial-up.adp
Why? It will turn out to be a good deal for them - someone else does all the upfront leg work and spending to put out a product - if they develop a hit, the big media companies can come in and sign them.
the bad news is if someone builds up sufficiant fanbase the record companies end up fighting over them. This means that unless there is collusion involved the band will get a far better deal than one coming from nothing and this will cut into the record companies profits.
afaict the bands that make the record companies the biggest profits are the big hits that are still locked into thier initial (signed when they were unknown) contracts.
ok so its not all that expensive per unit mass compared to some other comodities.
it is expensive per unit usable energy compared to other fuel sources (coal, natural gas) and the gap from other fuel sources is widening.
it is also expensive compared to its difficulty of extraction which makes the goverments with oil rescources very rich and powerfull.
and it also represents a huge transfer of money due to the sheer ammount of it extracted/used.
right now the same person has to create the good and bad original files then get people to sign the good one before he distributes the bad one.
and upon close inspection the good one would look very suspicious (lots of random garbage)
so its a break of sorts but nowhere near as bad as a preimage attack would be.
so china gets its oil slightly cheaper and the west slightly more expensive big deal
oil is a commodity, an expensive one but still a commodity. As such a single supplier cant really threaten a single customer (they can stop exporting thier oil altogether but that would hurt all oil customers as well as thier own pockets)
you could always have multiple seperate scams running at once after you built up some capital
The facts of the world here in fact have not changed. Only nomenclature has.
well what has actually happened is that nomenclature has caught up with the facts. The new(ish) fact is that there is a big belt of pluto like objects in similar orbits to pluto. The final straw came when it was discovered that pluto wasn't even the biggest in that belt.
It's also useful in education, because we ask our kids to learn the names of the planets, not every body that orbits the sun. There is really very little useful value in writing new textbooks here.
well you could say there is very little value in teaching our kids the makeup of the solar system at all. After all its not as though any significant number of people leave earth and the only bodies with significant impact on everyday life are the sun and moon.
what we have really discovered here is that pluto was not a one of a kind in a pretty unique orbit but part of a belt of very similar lumps of rock. School textbooks talk about the asteroid belt but not ceres in particular. Similarly they should talk about the kuiper belt but not pluto in particular.
Personally I reckon they should auction names rather than selling them at a flat rate.
for initial sales maybe and maybe renewals could be locked to the auction price but there is no way there should be an auction at renewal time.
domain name changes are very painfull for all involved and allowing anyone with the money to force someone else to undergo one on a whim is simply unacceptable and would tear apart the largely hobbyist structure of the net.
bad idea, one important design idea of domain names is they are supposed to be permanent (or as close to permanent as is reasonable) identifiers.
allowing a rich player to extort away a domain name from its long time user by outbidding them on registration fees would bring cahos to the net.
lets not replace an annoyance (cybersquatting) with something that will totally destroy the integrity of the naming system by allowing anyone to steal anyone elses name by paying a registration body more than the names owner can afford.
the REAL problem i'd imagine is that it takes very different design for a shuttle launch than for a big dumb rocket launch.
so if you plan to use the shuttle at all then the shuttle becomes a sticking point.
the government has a huge ability to both bribe (lots of money) and threaten (the government can very easilly make your life a misery for many years or of course they could just kill you and claim it was an accident).
punishing the companies is pointless for a descision that didn't involve the shareholders, it just hurts customers and shareholders alike while not really touching the execs. maybe the execs involved should be punished but i don't think even that is very fair given the pressure a government can introduce on an individual (any bribes should certainly be recovered though).
btw is the intel C compiler issue with amd chips (the one where non intel chips are sent down a slow codepath despite having the required feature bits to say they can use the fast one) still ongoing?
i should probablly clarify by accident before someone else does.
iirc it was discovered because an error in the then current understanding of orbital mechanics caused astronomers to belive there was a large object beyond neptune and pluto was discovered approximately where that large object would have been.
Heretofore a planet was (loosely) defined as a large mass in orbit around a star
the problem with that definition is that "large" is not well defined.
since the switch to the heliocentric model a planet has generally meant a large relatively unique (within the solar system at least) object that orbits the sun. Other objects orbiting the sun were given other names (e.g. asteriods, comets, kuiper belt objects)
the problem is like ceres we are discovering pluto is not a planet in that sense but part of a large belt of similar objects that while significant as a group don't really have anything to set them apart from each other. Like ceres it should probally be reclassified.
the reason its such an emotive issue is that pluto was discovered by accident long before the belt in which it lies was discovered. so it has been in grade school textbooks as a "planet" for a long time.
laptops are already allowed again iirc
i know its a diffiult balance, if you treat lusers like lusers then they will never switch unless forced, if you don't treat luesrs like lusers they get thier systems full of malware. Its a loose/loose choice.
what cds you wan't depends on the number of machines your installing and your ease of access to the internet during installations.
if you are planning to work away from an internet connection get the whole damn set of main CDs/DVDs.
if the machines don't have CD drives get the boot root and net-drivers floppies
if you are just doing one box and don't plan to use the cds after initial install get the netinst CD
if you wan't it to fit on a buisnesscard and don't care about ending up with an 486 optimised kernel (you can install an optimised one later from the net if you wan't) get buisnesscard.
otherwise get main CD 1.
installing stuff provided by your distro is easy, arguablly easier than installing stuff on windows (you don't have to go searching for download sources etc first)
the problem comes when you wan't something more custom. Something from outside the distros "walled garden" the distros are very resistant to universal packaging efforts like autopackage.
But since we are not using infinite number of decimal places, you are basically asking for a bigger float no?
NO
what matters in money is predictable behaviour according to specified rules, those rules are decimal based so the system needs to operate with a (generally fixed) decimal exponent.
bigger binary exponent floats will give a better approximation of 0.01 but the error is still there and will still add up over time.
I am positive that you will find more failures in people that are (a) very mobile and (b) have only one adapter.
yeah there are three things that stress connections afaict
1: use at full stretch (so any movement of anything gives heavy stress)
2: rough extraction from a tangle (so keep your cables untangled if possible, and if not possible at least try to be gentle with thier removal)
3: packing/unpacking, cables nearly always leave in directions that mean that they will have to either bend sharply to allow for use or bend sharply to allow the thing to be packed up reasonablly small.
which works fine until someone actually takes you to court and the court discovers that you used that reasoning and did a customer severe harm in the process.
thats what huge damages awards are really about, punishing corparations who use this line of reasoning.