The WLUG meeting is this evening ---------------------------------------------------------------------- Presenting at this meeting is Robert O'Callahan (http://weblogs.mozillazine.org/roc/), a member of the Ximian] desktop http://www.wlug.org.nz/Linux group at http://www.wlug.org.nz/Novell, responsible for http://www.wlug.org.nz/Mozilla/http://www.wlug.org.nz/Firefox ([MozillaFirefox) browser development and the integration of Firefox into Novell's Linux desktop product. The meeting will start at 7:30pm in the LITB (http://www.wlug.org.nz/LitB) room B.G.02 at WaikatoUniversity. Robert did his undergraduate study at the University of Auckland and has a PhD] from http://www.wlug.org.nz/Carnegie Mellon ([CMU) university. He has also previously worked at IBM Research. He made it known that he wanted a full time job working on Mozilla that he could do from New Zealand, and Novell hired him for their desktop team to work fulltime in Auckland on Mozilla improvements, including multi-column layout for web pages and improved SVG (scalable vector graphics) support. * Robert's blog (http://weblogs.mozillazine.org/roc/) * An article on Ben Goodger and Robert O'Callahan from the NZ Herald (http://www.nzherald.co.nz/index.cfm?c_id=11&ObjectID=10010075) Thanks to Douglas Harvey of the Systems & Technology Group at IBM] New Zealand and Chris Penney of http://www.wlug.org.nz/Novell for arranging this meeting. Beer and pizza will be provided compliments of [Value Added Distributors (http://www.vad.co.nz/), the principal distributor for IBM iSeries and pSeries equipment in New Zealand.
Hey Douglas, Even though I am writing from my work email address, this corespondance will probably carry on from my private email address as it is to do with my hosting services, so I may correspond from lindsay(a)wired.net.nz Can you please give a detail on the high level view of the IBM server range, ie. pSeries, xSeries, zSieries and so on and which path I should follow (and why). Details with the hosting requirements. Currently, I have built up a single P4 (3 GHz) which is a SuperMicro and has 2x 200 GB SATA HDDs, 2 Gigs of RAM and so on. I went down the SuperMicro path purely because it was seen as the cheapest solution since the whole project is a start up and I didn't want to throw huge capital into this project until I have a solid customer base and tread the waters carefully as so to speak. The OS path is Debian Linux, and other than dabbling with FreeBSD, I can see it remaining this way. I am open to other major distros if support is a question. For the current situation, the SuperMicro solution (from the server specifications point of view) has proven more than adequate. For future requirements, web/mail/dns servers will always me the front-end servers, and power usage, rack space usage, will always be a major consideration size those applications are not resource hungry. Backend servers are on the roadmap for housing the database which will provide information to the web pages and also to the customers, these servers will be firewalled and accessable to customers via secure web interface and/or vpn access, these servers I can forsee will be chosen with processing, memory, and storage as a major consideration and with one customer we're working with could end up huge. Which range of servers do you see is most appropiate for my requirements, considering overall bang for buck. I am looking at investing in IBM products rather than SuperMicro due to the fact that a friend of mine ships a lot of IBM kit, that is Craig Pollock from Alexis Infotech Management, and I've been convinced of the delivery times and the extended warranty which is very important. If you can help me here with the IBM product range, that would be really great. Best Regards, Lindsay Druett. Douglas Harvey wrote:
Hi
Thanks for letting Anatole Perry of VAD & myself attend the meeting last night, and put on beer & pizzas. Trust everyone whom enjoyed Robert's talk & the aftermatch function got home safely. :-)
Some urls for reference. http://www.vad.co.nz http://www.ibm.com/eserver/server/pseries http://www.ibm.com/linux http://www.ibm.com/developerworks/linux/power/ <http://www-128.ibm.com/developerworks/linux/power/>
University of Portland Linux on Power site http://upegr.up.edu:8080/
My contact details are; Douglas Harvey pSeries Technical Specialist mailto:dharvey(a)nz.ibm.com <mailto:REMOVE.dharvey(a)nz.ibm.com.REMOVE> http://www.ibm.com/ Ph 09 359 8409 if anyone wants to discuss the topics raised further.
Yeah, wrong address... :) Lindsay Druett wrote:
Hey Douglas,
Even though I am writing from my work email address, this corespondance will probably carry on from my private email address as it is to do with my hosting services, so I may correspond from lindsay(a)wired.net.nz
Can you please give a detail on the high level view of the IBM server range, ie. pSeries, xSeries, zSieries and so on and which path I should follow (and why).
Details with the hosting requirements.
Currently, I have built up a single P4 (3 GHz) which is a SuperMicro and has 2x 200 GB SATA HDDs, 2 Gigs of RAM and so on. I went down the SuperMicro path purely because it was seen as the cheapest solution since the whole project is a start up and I didn't want to throw huge capital into this project until I have a solid customer base and tread the waters carefully as so to speak.
The OS path is Debian Linux, and other than dabbling with FreeBSD, I can see it remaining this way. I am open to other major distros if support is a question. For the current situation, the SuperMicro solution (from the server specifications point of view) has proven more than adequate.
For future requirements, web/mail/dns servers will always me the front-end servers, and power usage, rack space usage, will always be a major consideration size those applications are not resource hungry. Backend servers are on the roadmap for housing the database which will provide information to the web pages and also to the customers, these servers will be firewalled and accessable to customers via secure web interface and/or vpn access, these servers I can forsee will be chosen with processing, memory, and storage as a major consideration and with one customer we're working with could end up huge.
Which range of servers do you see is most appropiate for my requirements, considering overall bang for buck.
I am looking at investing in IBM products rather than SuperMicro due to the fact that a friend of mine ships a lot of IBM kit, that is Craig Pollock from Alexis Infotech Management, and I've been convinced of the delivery times and the extended warranty which is very important.
If you can help me here with the IBM product range, that would be really great.
Best Regards, Lindsay Druett.
Douglas Harvey wrote:
Hi
Thanks for letting Anatole Perry of VAD & myself attend the meeting last night, and put on beer & pizzas. Trust everyone whom enjoyed Robert's talk & the aftermatch function got home safely. :-)
Some urls for reference. http://www.vad.co.nz http://www.ibm.com/eserver/server/pseries http://www.ibm.com/linux http://www.ibm.com/developerworks/linux/power/ <http://www-128.ibm.com/developerworks/linux/power/>
University of Portland Linux on Power site http://upegr.up.edu:8080/
My contact details are; Douglas Harvey pSeries Technical Specialist mailto:dharvey(a)nz.ibm.com <mailto:REMOVE.dharvey(a)nz.ibm.com.REMOVE> http://www.ibm.com/ Ph 09 359 8409 if anyone wants to discuss the topics raised further.
------------------------------------------------------------------------
_______________________________________________ wlug mailing list | wlug(a)list.waikato.ac.nz Unsubscribe: http://list.waikato.ac.nz/mailman/listinfo/wlug
Lindsay Druett wrote:
Yeah, wrong address... :)
At both ends; did you intend that to go to the list? :) -- Steve Balmer, CEO of Microsoft, once referred to LINUX as a cancer. Unsurprisingly, that's incorrect; LINUX was released on August 25th 1991 and is therefore a Virgo.
No, in fact, I'm going to bring it up with the committee as I'm not the only person that has been caught out this was where natually when you reply to someone, you hit reply, when you reply to a group, you hit reply all. When it comes to something confidential, I'm pretty pissed actually. This reply mungling is crap. zcat wrote:
Lindsay Druett wrote:
Yeah, wrong address... :)
At both ends; did you intend that to go to the list? :)
At 11:29 2/03/2005, Lindsay Druett wrote:
No, in fact, I'm going to bring it up with the committee as I'm not the only person that has been caught out this was where natually when you reply to someone, you hit reply, when you reply to a group, you hit reply all.
Do you confirm that the address in the To: header is correct when you're sending potentially confidential email? It's a "best practice" that I have established when sending any commercially sensitive correspondence, especially when I have document attachments. David
No, I like to think there is a human being behind my computer (rather than a robot) who doesn't spend his day double and triple checking stuff or is completely anal about things http://www.unicom.com/pw/reply-to-harmful.html David Hallett wrote:
At 11:29 2/03/2005, Lindsay Druett wrote:
No, in fact, I'm going to bring it up with the committee as I'm not the only person that has been caught out this was where natually when you reply to someone, you hit reply, when you reply to a group, you hit reply all.
Do you confirm that the address in the To: header is correct when you're sending potentially confidential email? It's a "best practice" that I have established when sending any commercially sensitive correspondence, especially when I have document attachments.
David
_______________________________________________ wlug mailing list | wlug(a)list.waikato.ac.nz Unsubscribe: http://list.waikato.ac.nz/mailman/listinfo/wlug
Lindsay Druett wrote:
No, I like to think there is a human being behind my computer (rather than a robot) who doesn't spend his day double and triple checking stuff or is completely anal about things
There are always counter-opinions: http://www.metasystema.org/essays/reply-to.mhtml There are wiki nodes on this too, although they don't do much other than link to these two sites.
Lindsay Druett wrote:
No, in fact, I'm going to bring it up with the committee as I'm not the only person that has been caught out this was where natually when you reply to someone, you hit reply, when you reply to a group, you hit reply all.
This comes up at least once with most mailing lists. In general I'm in favour of reply-to munging, because it encourages conversations to stay on the mailing list. Without reply-to munging, conversations divert off the mailing list very easily - especially with newer users. In general, the number of times one might want to reply privately to a person is much lower than the number of times one would want to reply to the list general, so setting reply-to munging makes sense. The problems I have with leaving the reply-to header intact are the aforementioned tendency to divert threads offlist, especially from newer users (whom this list is at least partially targetted at); and that it results in duplicate emails being delivered[1]. It also encourages people posting to the list and demanding "replies direct, as I don't read this list very often" - a fairly arrogant sort of attitude. [1] If I hit reply-all to an email from a list, it replies to the list and the original poster (OP). If someone else then hits reply-all to that email, it gets sent to me, the OP, and the list. If someone else then hits reply-all to that email, it gets to sent to the second responder, me, the OP, and the list, etc. While some MDAs are smart enough to suppress duplicate delivery, a lot of them arent. The problem actually lies with MUAs not really supporting the concept of mailing lists properly. Mutt, I believe, does it right, but I've not seen sensible support in any other MUA software. Something that let you do "Reply to List" and picked out the list email address,
[1] If I hit reply-all to an email from a list, it replies to the list and the original poster (OP). If someone else then hits reply-all to that email, it gets sent to me, the OP, and the list. If someone else then hits reply-all to that email, it gets to sent to the second responder, me, the OP, and the list, etc. While some MDAs are smart enough to suppress duplicate delivery, a lot of them arent.
Just to follow this point up, this point seems like it might have a bigger impact if we still lived in the days where you were changed $30/MB for email. Data is basically flatrate, so sending lots of emails doesn't matter any more. However, spam and virus scanning isn't cheap. I've seen threads on mailing lists CC'd to 6 lists members and the list itself. That's another 6 passes through spam and virus scanning processes. Unless spamassassin etc are smart enough to do duplicate suppression themselves - anyone have any idea if that's the case?
[1] If I hit reply-all to an email from a list, it replies to the list and the original poster (OP). If someone else then hits reply-all to that email, it gets sent to me, the OP, and the list. If someone else then hits reply-all to that email, it gets to sent to the second responder, me, the OP, and the list, etc. While some MDAs are smart enough to suppress duplicate delivery, a lot of them arent.
Recent mailman is smart enough to not deliver the duplicates, if the email address is subscribed to teh mailing list.
Lindsay Druett wrote:
No, in fact, I'm going to bring it up with the committee as I'm not the only person that has been caught out this was where natually when you reply to someone, you hit reply, when you reply to a group, you hit reply all.
When it comes to something confidential, I'm pretty pissed actually.
This reply mungling is crap.
As far as I can see this is perfectly sensible behaviour and email comes from a list, I hit reply it goes back to the list. The original email is from a mailing list to me, not from an individual to me, so if i reply to a post I expect to reply to the list not the individual
participants (8)
-
Daniel Lawson -
David Hallett -
Douglas Harvey -
Jason Drake -
Lindsay Druett -
Meeting Reminder bot -
Perry Lorier -
zcat