I have a related question, which is : "If I have an OS for that system, an OS whose vendor has licensed SCO source, does that license apply to my system when it runs linux"
I ask as I have a copy of win2K3 server in one vmware window, and could install Solaris x86 in another. Solaris is definitely SYSV legit, and Windows? well, MS have paid for the rights, my box came with windows preinstalled, so I still have those rights, no?
It certainly needs OWA, but I dont know about screen scraping.
I believe that Exchange2K also provides a WebDAV+XML interface to the server, which is more suited to programmatic access than pure screen scraping. Given a future edition of Exchange will provide SOAP access (so they say), it may be easier to talk to exchange in future -if MS want you to. They will probably emit key features or require NTLM authentication under the guise of 'security'...
heh, like only last week the big 'fix the RPC hole' patch email and voicemail went out, now we are being portscanned to see if we have fixed it.
The site managed linux distros -yup official debian derivatives- have a runlevel 3 script that apt-gets security patches from the server at boot time. So, once every three months they get updated...
well my desktop's Matrox card toasts the system if I pull out a monitor, and my laptop sporadically reboots with MS blaming the ATI card.
The latter irritates me as I cannot get an update of ATI (see the manufacturer, they say -but the manufacturer hasnt released an update since they shipped). I have better support on Xfree86, which makes it one of the few part of the laptop that is reliably supported on linux.
I think they've been holding off since that judge said they had to ship java with it...the longer they could put off the SP, the more chance of overturning the judgement.
Of course this means that if you get a box today you have to apply umpteen inadequately tested hotfixes, half of which cant be uninstalled, before your box is even vaguely secure, with reboots between every install. That really irritates me on those systems that I still have set up as dual boot...
There is also the OS panic uploads, with the corresponding web site, oca.microsoft.com.
Note how the site doesnt believe mozilla is up to date -if you replace IE with mozilla as your default web browser, you cannot see the post mortem analyses. You know, the ones that always blame the display driver rather than MS code...
Well, the LGPL in java issue means they may have a point were they to redist any of their code.
But if they have something against BSD licenses they'd better stop using Java 1.4 (Apache XML support), WebSphere (Xerces and Axis), and of course anything that implements the servlet2.2 or later spec, the reference implementation being Apache Tomcat. Which means no J2EE for them, then.
That is an interesting thought: it would stuff Linux dev overnight...the moment they turned off the servers or added the rule that everything committed became MS property.
But look on the bright side, the world would be free of Source Safe, which is possibly the worst SCM tool on the planet, worse even than zipping your source up into archive files (which gives you implicit labels that are accurate), or not doing any backup at all (which makes it obvious that your work will vanish when the disk drive breaks -SourceSafe pretends things are safe till they go horribly wrong)
1. normal apps only have a 'main' function that you call to run it
2. java JAR files are libraries with 0 or more classes that implement the static method main.
3. if you dont take the 'every java app is a library' view, then you can avoid any LGPL licenses that say derived libraries must be LGPL simply by providing a class with a main method.
As an example, apache Axis, the SOAP library I work on sometimes, has at least two entrypoints: org.apache.axis.client.AdminClient and org.apache.axis.client.HappyClient. Yet it is quite clearly a library, as server side you add it to your webapps lib dir to use the SOAP services.
the problem is that by importing lgpl code you are putting stuff in your class files that explicitly say 'I bind to this LGPL stuff'. Since every JAR is really a library, not just an app, then you are creating a new library which links into another LGPL lib. And therefore the LGPL also applies your JAR, which is 'a derivative library'
It is only if the public interfaces are not LPGL, and the binding is at runtime, that you are safe. For example, jboss is GPL, but as you use the J2EE API packages to talk to it, you are safe.
Re:And I suspect most of us feel the same way...
on
LGPL is Viral for Java
·
· Score: 2, Interesting
The apache interpretation has always been: every java app is a jar', so every app is a library, so for java, LGPL has the same semantics as GPL.
We now have official confirmation that this is the case, even if that is not what people who released LGPL java code intended. Maybe those people need to rerelease their code as GPL to formalise the outcome.
The effective result is that Apache Java projects are forbidden from linking to LGPL libraries, so we either go without the code or reimplement it -a loss either way.
you are correct -this is truly a wrong thing to do. I think they were smart enough to relegate windows to being the GUI for the front panel, and not do important things like keeping alcohol solutions below critical temperatures.
no doubt... to date the Grid is very java centric. Now maybe.NET could deliver a speedup, but the nice thing about Java is (a) the latest 1.4.2 JREs use the PIII SSE and P4 SSE2 register sets for better float and double performance, and (b) you can put some serious unix servers in the grid for bonus speed.
One thing Jim ignored is cost of software. Because MS effectively charge per-CPU for their system, you cannot afford to build a beowulf cluster on windows, let alone a full grid. So if MS do want to play in grid space, they need a way to price their platform so it makes economic sense. Didnt see that in the paper.
(nb, MS do clustering already, it is just focused at DBs and big IIS installations, and it costs big numbers)
yes, notebooks are the trouble spot. I have one of the new compaq designed xw8000 'workstations'; ordered it with rh7.3 preinstalled, and boosted it to rh9.0 the moment I got it -then downloaded the nvidia drivers for decent 3d performance on xscreensaver.
Which leaves the laptops. The problem here is that (a) they all come from compal or quanta in taipei, and (b) you cant pick and choose cards to get proper display, modem and wlan. The PC vendors could qualify them for linux under commercial pressure, but without that pressure they are not going to bother. And frankly, the./ community are not the premium purchasers of kit that can justify such a move. Either its because the corporates want it, or because the vendors discover they can do ultra low cost laptops with linux that sell in EU and north america.
There is a fascinating online book on the subject,
Secret Fallout. What is interesting is not just the statistics (increase in child mortalities and failed pregnancies after nuclear tests), but all the institutional resistance. The Atomic energy industry didnt want to believe the resuls, and the (primarily government) institutions fought its publication and argued every step of the way.
These are of course, the same people saying these backscatter x-rays are safe. If they arent, well, a lot of people are going to have been exposed before that becomes clear.
I am not outstandingly worried as (a) the exposure from cosmic rays of any over-the-pole flight is bound to be worse, and (b) back in '85 I was caught out in rain that turned out to be from chernobyl.That was in Scotland, where some places are still so radioactive that they have to
decontaminate sheep before sale for human consumption.
Jboss tun a profit on two things: consultancy and docs. They most definitely do not release their docs on any GPL-like terms, and they ship jboss with nearly no docs at all.
What irritates me about that is that their distros ship without all the docs from Apache covering underlying components (Axis, Tomcat &c), which are freely available. There is no justification for that, other than it would make jboss look bad,
I actually use formal methods when proving that threads syncrhonize/dont deadlock: its a dog to test such things and a bit of mathematics saves a lot of time. The rest of the time: JUnit and CPPUnit tests. The nice thing about tests is that you can automate regression testing -I've never heard of automated regression proofs...
oh, I didnt forget cromwell; the Restoration was what I was referring to in 'we blew our chance at a republic. Or as parliament said after executing charles 1, "the office of the king in this nation is unnecessary, burdensome and dangerous to the liberty, society and public interest of the people.""
Given that the house of lords is sort of no more -and that was the real instantiation of the aristocracy in the goverment, it can only be a matter of time before we have to take another close look at royalty.
But till then, whoever in charge can veto things. It would create a major constitutional crisis, but they could do it. It'd be fun to watch:)
That's the US; in the UK we are 'subjects of the crown'; the govenment reports to whoever is King or Queen at the time.
Yes, I know it makes as much sense as a 64-way NT server, but we blew our chance to be a republic back in the seventeeth century, after the English Civil War.
Given I am going to get my employer to pay $400 for VMware, so that I can run all my winapps, I can say that paying for stuff can be justified. But that is a special case - I am paying not just to get outlook and the rest of office to work, but for the ability to have a disk full of different OS images and then write and debug complex windows apps/device drivers and not even have to blink when an OS image bluescreens on me.
Otherwise, well, $600 is a lot for any OS these days, given you can buy a reasonable PC for that much.
I see your point, but what irritates me w/ Evo is that it repeats so many of the UI mistakes of outlook.
Case in point: why doesnt the calendar highlight *in* the day view which day is today, when is 'now' (i.e. a coloured line) and when is past (i.e. a darker gray). If you are trying to make an appointment over the phone, it is too easy to accidentally make that appointment in the past, because the GUI doesnt highlight past/present/future properly.
At least with Evo I can fix such details; with outlook you have to be grateful for what you get.
This may have been the original plan, but look at it now: the cellphone vendors get to charge extra roaming/long distance fees when you use your phone outside the 'home area', and double charge on intra-network calls. That and the round up to the next minute plan all brings in bonus $.
I am so looking forward to getting a decent phone and decent service when I return to europe. -steve
Axis includes TZ info if you send a java.util.Calendar object; it only assumes Zulu for java.util.Date instances. So use Calendar instancess everywhere (which is what it currently does when mapping WSDL to java) and you do retain the info.
But.NET assumes local timezone for everything, mapping it down to the.NET wrapper round time_t, dropping TZ info and generally making a mess. That is why I have to use time_t, and document the fact that this is always in UTC in the service specs. An ugly hack, I know.
One issue the SoapBuilders have with the XML schema is there is no way to say 'unknown TZ' in the language, so you have to say *something*.
Slashdot Mirror
I have a related question, which is : "If I have an OS for that system, an OS whose vendor has licensed SCO source, does that license apply to my system when it runs linux"
I ask as I have a copy of win2K3 server in one vmware window, and could install Solaris x86 in another. Solaris is definitely SYSV legit, and Windows? well, MS have paid for the rights, my box came with windows preinstalled, so I still have those rights, no?
It certainly needs OWA, but I dont know about screen scraping.
I believe that Exchange2K also provides a WebDAV+XML interface to the server, which is more suited to programmatic access than pure screen scraping. Given a future edition of Exchange will provide SOAP access (so they say), it may be easier to talk to exchange in future -if MS want you to. They will probably emit key features or require NTLM authentication under the guise of 'security'...
Because it is only the business end users who dont have many, many lawyers who are going to pay up.
:
If they go after the distributors (they will), the distributors will say 'see you in court' and SCO wont get any cash flow for years, if any.
One thing I havent seen covered is
If you think you really need a SYSV license to go legit, why not get a Solaris/x86 license, instead of the SCO one?
That way you get a reasonable OS and it costs you less. Plus it denies SCO money yet should put you in a good defensible position.
If SCO dont even have the ability to check on solaris licenses, you can just say 'we have solaris licenses, go away' and be done with it...
heh, like only last week the big 'fix the RPC hole' patch email and voicemail went out, now we are being portscanned to see if we have fixed it.
The site managed linux distros -yup official debian derivatives- have a runlevel 3 script that apt-gets security patches from the server at boot time. So, once every three months they get updated...
well my desktop's Matrox card toasts the system if I pull out a monitor, and my laptop sporadically reboots with MS blaming the ATI card.
The latter irritates me as I cannot get an update of ATI (see the manufacturer, they say -but the manufacturer hasnt released an update since they shipped). I have better support on Xfree86, which makes it one of the few part of the laptop that is reliably supported on linux.
I think they've been holding off since that judge said they had to ship java with it...the longer they could put off the SP, the more chance of overturning the judgement.
Of course this means that if you get a box today you have to apply umpteen inadequately tested hotfixes, half of which cant be uninstalled, before your box is even vaguely secure, with reboots between every install. That really irritates me on those systems that I still have set up as dual boot...
There is also the OS panic uploads, with the corresponding web site, oca.microsoft.com. Note how the site doesnt believe mozilla is up to date -if you replace IE with mozilla as your default web browser, you cannot see the post mortem analyses. You know, the ones that always blame the display driver rather than MS code...
Well, the LGPL in java issue means they may have a point were they to redist any of their code.
But if they have something against BSD licenses they'd better stop using Java 1.4 (Apache XML support), WebSphere (Xerces and Axis), and of course anything that implements the servlet2.2 or later spec, the reference implementation being Apache Tomcat. Which means no J2EE for them, then.
That is an interesting thought: it would stuff Linux dev overnight...the moment they turned off the servers or added the rule that everything committed became MS property.
But look on the bright side, the world would be free of Source Safe, which is possibly the worst SCM tool on the planet, worse even than zipping your source up into archive files (which gives you implicit labels that are accurate), or not doing any backup at all (which makes it obvious that your work will vanish when the disk drive breaks -SourceSafe pretends things are safe till they go horribly wrong)
1. normal apps only have a 'main' function that you call to run it
2. java JAR files are libraries with 0 or more classes that implement the static method main.
3. if you dont take the 'every java app is a library' view, then you can avoid any LGPL licenses that say derived libraries must be LGPL simply by providing a class with a main method.
As an example, apache Axis, the SOAP library I work on sometimes, has at least two entrypoints: org.apache.axis.client.AdminClient and org.apache.axis.client.HappyClient. Yet it is quite clearly a library, as server side you add it to your webapps lib dir to use the SOAP services.
the problem is that by importing lgpl code you are putting stuff in your class files that explicitly say 'I bind to this LGPL stuff'. Since every JAR is really a library, not just an app, then you are creating a new library which links into another LGPL lib. And therefore the LGPL also applies your JAR, which is 'a derivative library'
It is only if the public interfaces are not LPGL, and the binding is at runtime, that you are safe. For example, jboss is GPL, but as you use the J2EE API packages to talk to it, you are safe.
The apache interpretation has always been: every java app is a jar', so every app is a library, so for java, LGPL has the same semantics as GPL.
We now have official confirmation that this is the case, even if that is not what people who released LGPL java code intended. Maybe those people need to rerelease their code as GPL to formalise the outcome.
The effective result is that Apache Java projects are forbidden from linking to LGPL libraries, so we either go without the code or reimplement it -a loss either way.
you are correct -this is truly a wrong thing to do. I think they were smart enough to relegate windows to being the GUI for the front panel, and not do important things like keeping alcohol solutions below critical temperatures.
I saw a refrigerator in Fryes that had an embedded win9x box and monitoring doing added value stuff: tv, music, message board, web browsing.
I know it is win9x as for a test we flipped the power and scandisc came up on reboot...
no doubt ... to date the Grid is very java centric. Now maybe .NET could deliver a speedup, but the nice thing about Java is (a) the latest 1.4.2 JREs use the PIII SSE and P4 SSE2 register sets for better float and double performance, and (b) you can put some serious unix servers in the grid for bonus speed.
One thing Jim ignored is cost of software. Because MS effectively charge per-CPU for their system, you cannot afford to build a beowulf cluster on windows, let alone a full grid. So if MS do want to play in grid space, they need a way to price their platform so it makes economic sense. Didnt see that in the paper.
(nb, MS do clustering already, it is just focused at DBs and big IIS installations, and it costs big numbers)
yes, notebooks are the trouble spot. I have one of the new compaq designed xw8000 'workstations'; ordered it with rh7.3 preinstalled, and boosted it to rh9.0 the moment I got it -then downloaded the nvidia drivers for decent 3d performance on xscreensaver.
./ community are not the premium purchasers of kit that can justify such a move. Either its because the corporates want it, or because the vendors discover they can do ultra low cost laptops with linux that sell in EU and north america.
Which leaves the laptops. The problem here is that (a) they all come from compal or quanta in taipei, and (b) you cant pick and choose cards to get proper display, modem and wlan. The PC vendors could qualify them for linux under commercial pressure, but without that pressure they are not going to bother. And frankly, the
These are of course, the same people saying these backscatter x-rays are safe. If they arent, well, a lot of people are going to have been exposed before that becomes clear.
I am not outstandingly worried as (a) the exposure from cosmic rays of any over-the-pole flight is bound to be worse, and (b) back in '85 I was caught out in rain that turned out to be from chernobyl.That was in Scotland, where some places are still so radioactive that they have to decontaminate sheep before sale for human consumption.
Jboss tun a profit on two things: consultancy and docs. They most definitely do not release their docs on any GPL-like terms, and they ship jboss with nearly no docs at all.
What irritates me about that is that their distros ship without all the docs from Apache covering underlying components (Axis, Tomcat &c), which are freely available. There is no justification for that, other than it would make jboss look bad,
Good point about market size.
I actually use formal methods when proving that threads syncrhonize/dont deadlock: its a dog to test such things and a bit of mathematics saves a lot of time. The rest of the time: JUnit and CPPUnit tests. The nice thing about tests is that you can automate regression testing -I've never heard of automated regression proofs...
oh, I didnt forget cromwell; the Restoration was what I was referring to in 'we blew our chance at a republic. Or as parliament said after executing charles 1, "the office of the king in this nation is unnecessary, burdensome and dangerous to the liberty, society and public interest of the people.""
:)
Given that the house of lords is sort of no more -and that was the real instantiation of the aristocracy in the goverment, it can only be a matter of time before we have to take another close look at royalty.
But till then, whoever in charge can veto things. It would create a major constitutional crisis, but they could do it. It'd be fun to watch
That's the US; in the UK we are 'subjects of the crown'; the govenment reports to whoever is King or Queen at the time.
Yes, I know it makes as much sense as a 64-way NT server, but we blew our chance to be a republic back in the seventeeth century, after the English Civil War.
Given I am going to get my employer to pay $400 for VMware, so that I can run all my winapps, I can say that paying for stuff can be justified. But that is a special case - I am paying not just to get outlook and the rest of office to work, but for the ability to have a disk full of different OS images and then write and debug complex windows apps/device drivers and not even have to blink when an OS image bluescreens on me.
Otherwise, well, $600 is a lot for any OS these days, given you can buy a reasonable PC for that much.
I see your point, but what irritates me w/ Evo is that it repeats so many of the UI mistakes of outlook.
Case in point: why doesnt the calendar highlight *in* the day view which day is today, when is 'now' (i.e. a coloured line) and when is past (i.e. a darker gray). If you are trying to make an appointment over the phone, it is too easy to accidentally make that appointment in the past, because the GUI doesnt highlight past/present/future properly.
At least with Evo I can fix such details; with outlook you have to be grateful for what you get.
This may have been the original plan, but look at it now: the cellphone vendors get to charge extra roaming/long distance fees when you use your phone outside the 'home area', and double charge on intra-network calls. That and the round up to the next minute plan all brings in bonus $.
I am so looking forward to getting a decent phone and decent service when I return to europe.
-steve
Axis includes TZ info if you send a java.util.Calendar object; it only assumes Zulu for java.util.Date instances. So use Calendar instancess everywhere (which is what it currently does when mapping WSDL to java) and you do retain the info.
.NET assumes local timezone for everything, mapping it down to the .NET wrapper round time_t, dropping TZ info and generally making a mess. That is why I have to use time_t, and document the fact that this is always in UTC in the service specs. An ugly hack, I know.
But
One issue the SoapBuilders have with the XML schema is there is no way to say 'unknown TZ' in the language, so you have to say *something*.