Did you really try cyrus or did you just dump it because it looks similar to maildb?
cyrus really has some interesting features and is way faster than mbox:
- full IMAP-4.1 complicance with multi access - ACL - Quota - sieve support - hard link support for multiple recipients (yes, this means sending a 10 MB file to all local users will take 10 MB disk space on the mail server).
Very impressive article.
However, unfortunately, the author got the EUI-64 algorithm wrong. Before inserting the fffe in the middle of the MAC address, the universal/local bit has to be inverted. So 01:23:45:67:89:ab will give you as host part of the IPv6 address: 0323:45ff:fe67:89ab
http://www.linas.org/linux/i370.html has some details about the beginning of the S/390
port. The port can even be run on the Hercules 370/390 simulator, although it might run a bit slow;-).
Also, the S/390 code is only really useful with binary-only drivers for lcs (3172, old OSA) and qdio (GB OSA). Until these drivers are available in full source, i wouldn't call the port complete;-)
SuSE 6.4 is announced to be IPv6 ready. PLD (http://www.pld.org.pl/) already is. The current rawhide contains at least basic IPv6 functionality (both ping and traceroute are there).
That's right, but you will still need a router between the ISP line and your local net. And if you run Linux 2.4.0-test2 with the latest netfilter on this router, you even have a working IPv6 firewall:-). Just hope future Linux distributions with IPv6 (SuSE already announced IPv6 for 6.4, PLD ships IPv6 for a while) and 2.4.0 kernels will ship with a predefined simple firewall included.
On the other hand, NAT usually prevents using end-to-end authentication with IPSEC (also part of IPv6), as a changed IP header will invalidate the signature over the IP packet.
Re:Clearing some things up...
on
WAP Under Fire
·
· Score: 1
> The second article has been misleadingly > described. "IcesTorm-I sent us an IETF document > criticizing the format". This IS NOT an IETF > document, rather a post to an IETF mailing list. > The points contained in it about WAP not being a > truly open format are valid. The technical > arguements are a mixed bag - yes, some of the > wheels reinvented by the WAPForum definetly have > corners.:-)
Sure, however there have been a lot of response to this post from various people which are very active in the IETF.
The main concern from them was that apparently the WAPForum members didn't learn the lessons of the past and again try to invent yet another network protocol which has several flaws (patent issues and the broken encryption stuff, as well as the required "gateway" service between the WAP protocol suite and IP). One contributor even mentioned, that current mobile phones have a larger bandwith than the internet backbone of the early days of TCP/IP!
The good thing about most Free Software projects is that they are a community effort and no single party holds the copyright on any core part of the project. This guarantees that nobody has the possibility to block development or make special license agreements with certain parties giving them an advantage over others.
Qt does definitely not fall into the above category. A single entity (TT) owns all the code and even some special agreements with other companies exist (like Borland, IIRC). Sure, it is pure speculation if Borland has some special agreements with TT to give them early access to any new feature being added to Qt, but it is still possible.
So IMHO, making Qt (or any other "only open source, but not community based" product) part of the core of a community based free software project (like Linux), this would compromise the advantage of the project being a fair and open playground for everybody.
This doesn't mean that Open Source projects from companies and in particular Qt are bad things, but we should be sure to double think about the kind of problems we might run into when relying a larger project (like KDE) on them.
Many big banks in Germany use SAS on both mainframes (S/390) and Unix. Having SAS on Linux really helps the developers of SAS solutions as most of them don't have their own workstation to develop their software on. And development on NT and CPORTing the stuff to Unix is no good solution, either...
I really hope that SAS not only ports the other Unix versions over to Linux, but also provide the client parts like Enterprise reporter, database miner or IT Service Vision (former CPE), as well.
> counting on the fact that 80% of their audience > are IE users, and the remaining 20% are used to > meing marginalized.
Depending on the "features" being used, the 20% might easily raise to maybe 60% without the author even knowing. Reason is that most (big) companies firewalls do some filtering based on content. E.g. i know of no German bank which doesn't block Active X... And on our non-commercial site www.scram.de, almost 40% of all accesses are from within big companies (seems their employees have most spare time:) )
The MAC address being part of the IPv6 address is NOT mandatory. It may just be used for autosetup (just like the MAC address being part of an IPX address, as well). It is never used for routing or address resolutions anywhere. Neighbor solicitation and neighbor advertisements do the resolution in the local network and take over the rule or ARP from IPv4.
Using the network card MAC address as part of the IPv6 address is only one way of setting up the global IPv6 addresses (it's unmanaged autoconfiguration used by rtadvd). Alternatives are manual configuration or using DHCP with IPv6 extension.
One of the leading German EDA companies, CadSoft has already released a Linux version of their PCB CAD package which is actively marketed the same way as the DOS and Windows 95/NT versions. So, apparently, there is a market.
For Windows 98, the argument of device drivers being available might be true (but just you can't sell any hardware without Windows 95/98 device drivers). For Windows NT, however, i don't see more device drivers than Linux, the exact opposite seems to be the case. And finally in terms of supported file systems, i still have to find a system which has wider support than Linux.
I just wonder what part of NetWare Novell plans to release sources for. I don't believe the sources to yet another operating system will be a great benefit for the OSS community. However, the source to the NetWare network components (netware protocol, user database, directory services etc) could be very interesting for making these protocols the defacto (or even IETF approved) standard instead of the SMB protocol with all its flaws for hererogenous environments.
I probably wouldn't switch to writing NLMs, just because NetWare is free, but i would switch to the NetWare protocol, once the reference implementation is free.
Make the source independent on the jokes and put the jokes into a seperate configuration file. Then distribute the source with a non-offending sample configuration.
IMHO, making offending jokes about former co-workers isn't exaclty what i call good style, anyways...
The fine part about proprietary software is that you only get it for mainstream platforms, i.e. Windows, Mac and maybe GNU/Linux/i386 (and in case of RA, SGI)
I really hope the FreeAmp MP3/RTP project will be a usable free alternative to this RA crap.
I would really love to see Linux support for these boxes (i.e. Intel multibus + the strange IO they are using). Sure, MX300 are not the fastest machines (although they do run with Pentium processors), but they have proved to be really reliable (they even have 2 power supplies in case one fails).
Slashdot Mirror
Did you really try cyrus or did you just dump it because it looks similar to maildb?
cyrus really has some interesting features and is way faster than mbox:
- full IMAP-4.1 complicance with multi access
- ACL
- Quota
- sieve support
- hard link support for multiple recipients (yes, this means sending a 10 MB file to all local users will take 10 MB disk space on the mail server).
And it proved to be very reliable.
--jochen
Very impressive article.
However, unfortunately, the author got the EUI-64 algorithm wrong. Before inserting the fffe in the middle of the MAC address, the universal/local bit has to be inverted. So 01:23:45:67:89:ab will give you as host part of the IPv6 address: 0323:45ff:fe67:89ab
--jochen
In Europe, the RIPE would have rejected this design.
--jochen
Also, the S/390 code is only really useful with binary-only drivers for lcs (3172, old OSA) and qdio (GB OSA). Until these drivers are available in full source, i wouldn't call the port complete ;-)
--jochen
Does that mean that IBM will put these major parts of Linux/S390 under an OpenSource license, as well?
-- jochen
- zebra
- gated
- VRRP (the IETF equivalent of HSRP)
-- jochenBut Linux is coming closer...
SuSE 6.4 is announced to be IPv6 ready.
PLD (http://www.pld.org.pl/) already is.
The current rawhide contains at least basic IPv6 functionality (both ping and traceroute are there).
That's right, but you will still need a router between the ISP line and your local net. And if you run Linux 2.4.0-test2 with the latest netfilter on this router, you even have a working IPv6 firewall :-). Just hope future Linux distributions with IPv6 (SuSE already announced IPv6 for 6.4, PLD ships IPv6 for a while) and 2.4.0 kernels will ship with a predefined simple firewall included.
On the other hand, NAT usually prevents using end-to-end authentication with IPSEC (also part of IPv6), as a changed IP header will invalidate the signature over the IP packet.
> The second article has been misleadingly :-)
> described. "IcesTorm-I sent us an IETF document
> criticizing the format". This IS NOT an IETF
> document, rather a post to an IETF mailing list.
> The points contained in it about WAP not being a
> truly open format are valid. The technical
> arguements are a mixed bag - yes, some of the
> wheels reinvented by the WAPForum definetly have
> corners.
Sure, however there have been a lot of response to this post from various people which are very active in the IETF.
The main concern from them was that apparently the WAPForum members didn't learn the lessons of the past and again try to invent yet another network protocol which has several flaws (patent issues and the broken encryption stuff, as well as the required "gateway" service between the WAP protocol suite and IP). One contributor even mentioned, that current mobile phones have a larger bandwith than the internet backbone of the early days of TCP/IP!
-- jochen
IMHO, that's exactly the point.
The good thing about most Free Software projects is that they are a community effort and no single party holds the copyright on any core part of the project. This guarantees that nobody has the possibility to block development or make special license agreements with certain parties giving them an advantage over others.
Qt does definitely not fall into the above category. A single entity (TT) owns all the code and even some special agreements with other companies exist (like Borland, IIRC). Sure, it is pure speculation if Borland has some special agreements with TT to give them early access to any new feature being added to Qt, but it is still possible.
So IMHO, making Qt (or any other "only open source, but not community based" product) part of the core of a community based free software project (like Linux), this would compromise the advantage of the project being a fair and open playground for everybody.
This doesn't mean that Open Source projects from companies and in particular Qt are bad things, but we should be sure to double think about the kind of problems we might run into when relying a larger project (like KDE) on them.
-- Jochen
Many big banks in Germany use SAS on both mainframes (S/390) and Unix. Having SAS on Linux really helps the developers of SAS solutions as most of them don't have their own workstation to develop their software on. And development on NT and CPORTing the stuff to Unix is no good solution, either...
I really hope that SAS not only ports the other Unix versions over to Linux, but also provide the client parts like Enterprise reporter, database miner or IT Service Vision (former CPE), as well.
-- Jochen
> counting on the fact that 80% of their audience
:) )
> are IE users, and the remaining 20% are used to
> meing marginalized.
Depending on the "features" being used, the 20% might easily raise to maybe 60% without the author even knowing. Reason is that most (big) companies firewalls do some filtering based on content. E.g. i know of no German bank which doesn't block Active X... And on our non-commercial site www.scram.de, almost 40% of all accesses are from within big companies (seems their employees have most spare time
-- Jochen
The MAC address being part of the IPv6 address is NOT mandatory. It may just be used for autosetup (just like the MAC address being part of an IPX address, as well). It is never used for routing or address resolutions anywhere. Neighbor solicitation and neighbor advertisements do the resolution in the local network and take over the rule or ARP from IPv4.
-- Jochen
Using the network card MAC address as part of the IPv6 address is only one way of setting up the global IPv6 addresses (it's unmanaged autoconfiguration used by rtadvd). Alternatives are manual configuration or using DHCP with IPv6 extension.
-- Jochen
It looks like the developers of this distribution are not aware of SAS.
There even is an effort to get SAS ported to Linux.
-- Jochen
One of the leading German EDA companies, CadSoft has already released a Linux version of their PCB CAD package which is actively marketed the same way as the DOS and Windows 95/NT versions. So, apparently, there is a market.
-- Jochen
Hi,
:)
I just want to mention The Linux Token Ring Project. Here you may find the latest and greatest TR drivers and related information
-- Jochen
I'm really looking forward for tomorrow, so this stops again :)
-- Jochen
For Windows 98, the argument of device drivers being available might be true (but just you can't sell any hardware without Windows 95/98 device drivers). For Windows NT, however, i don't see more device drivers than Linux, the exact opposite seems to be the case. And finally in terms of
supported file systems, i still have to find a system which has wider support than Linux.
-- Jochen
I just wonder what part of NetWare Novell plans to release sources for. I don't believe the sources to yet another operating system will be a great benefit for the OSS community. However, the source to the NetWare network components (netware protocol, user database, directory services etc) could be very interesting for making these protocols the defacto (or even IETF approved) standard instead of the SMB protocol with all its
flaws for hererogenous environments.
I probably wouldn't switch to writing NLMs, just because NetWare is free, but i would switch to the NetWare protocol, once the reference implementation is free.
-- Jochen
So we should form a working group and submit the :)
first drafts then
-- Jochen
Make the source independent on the jokes and
put the jokes into a seperate configuration file.
Then distribute the source with a non-offending
sample configuration.
IMHO, making offending jokes about former
co-workers isn't exaclty what i call good style,
anyways...
-- Jochen
And the thread is still growing...
-- Jochen
Anyone having a player for:
- HP/UX 10.20
- GNU/Linux/68K
- GNU/Linux/AXP
- NetBSD/68K
The fine part about proprietary software is that
you only get it for mainstream platforms, i.e.
Windows, Mac and maybe GNU/Linux/i386 (and in
case of RA, SGI)
I really hope the FreeAmp MP3/RTP project will
be a usable free alternative to this RA crap.
-- Jochen
Are you talking about the SNI MX300i series?
I would really love to see Linux support for these
boxes (i.e. Intel multibus + the strange IO they
are using). Sure, MX300 are not the fastest
machines (although they do run with Pentium
processors), but they have proved to be really
reliable (they even have 2 power supplies in case
one fails).
-- Jochen