Think of this as a motion as one might move at a meeting. Discussion and refinement are in order.
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I am not on the Fedora users' list because of the volume of email, though I do use Fedora Core and might usefully contribute to FC in that way.
Similarly, I'm no longer on the OpenSUSE list, same story.
The time has come when I must, again, reduce the volume of email I see, because it's clogging my modem.
I use CentOS4 and plan to use CentOS5. I am interested in email for those, but not for CentOS 3 (I have no systems) nor CentOS 2 (I have one RHL 7.3 system that thinks it's CentOS 2, but it's in maintenance mode, and anything I do I will do alone).
I don't know how much email I would eliminate by dropping email for Centos < C4, but when I move to CentOS 5 I will lose interest in C4 and then there will be savings.
If this list were split into one for each release, then subscribers could choose which email they see. At present, it's all or none, and neither suits me.
If this list is split into four, then I expect the transition method would be to subscribe everyone on this list to the new four.
At some point, this would become read-only (nobody posts), or maybe all mail for this goes (via a filter to fix the headers) to all the others.
By "fix the headers" I mean "do whatever it takes to ensure replies go to the list the subscriber is replying to."
I'm sure this transition arrangement is imperfect; my objectives are to encourage CentOS to have the separate lists, and to ensure that the transition is fairly easy for users so we don't lose lots of subscribers.
What do others think?
On Tuesday, May 01, 2007 9:24 AM +0800 John Summerfield debian@herakles.homelinux.org wrote:
I don't know how much email I would eliminate by dropping email for Centos < C4, but when I move to CentOS 5 I will lose interest in C4 and then there will be savings.
How much mail is release-specific? I would guess such threads would be mostly about installation, so it might be better to split off a -install list. Traffic about getting drivers to work might also be redirected to such a list.
Runtime packages should be pretty common among all releases and might as well be in -users.
You might also consider reading the list with a news reader through gmane to avoid the download cost.
Kenneth Porter wrote:
On Tuesday, May 01, 2007 9:24 AM +0800 John Summerfield debian@herakles.homelinux.org wrote:
I don't know how much email I would eliminate by dropping email for Centos < C4, but when I move to CentOS 5 I will lose interest in C4 and then there will be savings.
How much mail is release-specific? I would guess such threads would be mostly about installation, so it might be better to split off a -install list. Traffic about getting drivers to work might also be redirected to such a list.
Installation and remastering 4 is very different from 5, following the intro of yum into Anaconda.
Applications to do stuff tend to change wu-imapd is out, dovecot and cyrus are in. Pine is out, mutt is in. Networking configuration's changed. lprNG is out, CUPS is in. Lotsa stuff changes. Have a look at the release notes, see what's expunged and what's deprecated.
Runtime packages should be pretty common among all releases and might as well be in -users.
You might also consider reading the list with a news reader through gmane to avoid the download cost.
That would be worse: [summer@bilby ~]$ ping -c4 terad.net PING terad.net (203.15.140.104) 56(84) bytes of data. 64 bytes from terad.net (203.15.140.104): icmp_seq=0 ttl=247 time=1116 ms 64 bytes from terad.net (203.15.140.104): icmp_seq=1 ttl=247 time=1345 ms 64 bytes from terad.net (203.15.140.104): icmp_seq=2 ttl=247 time=366 ms 64 bytes from terad.net (203.15.140.104): icmp_seq=3 ttl=247 time=1504 ms
--- terad.net ping statistics --- 4 packets transmitted, 4 received, 0% packet loss, time 5380ms rtt min/avg/max/mdev = 366.190/1083.366/1504.692/436.421 ms, pipe 3 [summer@bilby ~]$
I'm pulling a gbyte/month through a modem. Interactive through that modem is largely out.
John Summerfield wrote:
rtt min/avg/max/mdev = 366.190/1083.366/1504.692/436.421 ms, pipe 3 ... I'm pulling a gbyte/month through a modem. Interactive through that modem is largely out.
ouch. I'm averaging about 10-20GB each way a month over a 768k sdsl at home, and manage to maintain 10-15mS pings to my ISP's peering gateways most of the time. I can't even imagine 1 second pings
John Summerfield wrote:
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I totally agree John - as CentOS gets more popular i too have found myself using 'Mark folder as Read' too often.
If this list were split into one for each release, then subscribers could choose which email they see. At present, it's all or none, and neither suits me.
My worry would be the increases in cross-posting, and maybe the need for centos-general, but i guess these can mostly be posted to a CentOS(max_version) list.
A point that may mitigate the need to act now: I have seen the volume of mail surge about t-2_months before the launch of v5 - and i wasnt here for the launch of 4.0 -> 4.4, but would you not expect a lot of noise to dissipate soon now we are t+2_weeks?
Thats my 2c worth.
MrKiwi
Quoting MrKiwi mrkiwi@gmail.com:
John Summerfield wrote:
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I totally agree John - as CentOS gets more popular i too have found myself using 'Mark folder as Read' too often.
This is a good thing. I've just left the Fedora Core mailing lists (users & devel) and the traffic didn't bother me at all. It's a great thing to have a busy list - as it means that there are more eyeballs looking at just about everything. I browse a lot of things that are posted and although I don't reply to a lot, if I have a suggestion and have time to write it out, I will.
If this list were split into one for each release, then subscribers could choose which email they see. At present, it's all or none, and neither suits me.
My worry would be the increases in cross-posting, and maybe the need for centos-general, but i guess these can mostly be posted to a CentOS(max_version) list.
Splitting by version is usually a bad idea. It causes things to split, people post the wrong version information to the wrong lists and it becomes a PITA. Hell, I've noticed enough people whinge and moan about top posting... Imagine that on someone posting a v5 question to a v4 list!
A point that may mitigate the need to act now: I have seen the volume of mail surge about t-2_months before the launch of v5 - and i wasnt here for the launch of 4.0 -> 4.4, but would you not expect a lot of noise to dissipate soon now we are t+2_weeks?
This depends on your definition of noise. As long as the posts are CentOS related, I wouldn't care if there was 100 or 1000 posts per day. I'll do the same as what I do now. Skim through the subjects (a hint - make your subject useful and more eyeballs will read your post!). What I do care about is people posting random junk and making the signal to noise ratio go down the crapper.
Thats my 2c worth.
I'll see your 2c, and raise you 2c.
Steven Haigh wrote:
Quoting MrKiwi mrkiwi@gmail.com:
John Summerfield wrote:
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I totally agree John - as CentOS gets more popular i too have found myself using 'Mark folder as Read' too often.
This is a good thing. I've just left the Fedora Core mailing lists (users & devel) and the traffic didn't bother me at all. It's a great
Not everyone can have such a fine Internet connexion.
Splitting by version is usually a bad idea. It causes things to split, people post the wrong version information to the wrong lists and it becomes a PITA. Hell, I've noticed enough people whinge and moan about top posting... Imagine that on someone posting a v5 question to a v4 list!
I thought it worked very well indeed with Red Hat Linux. I note that Red Hat continues with that plan with RHEL.
A point that may mitigate the need to act now: I have seen the volume of mail surge about t-2_months before the launch of v5 - and i wasnt here for the launch of 4.0 -> 4.4, but would you not expect a lot of noise to dissipate soon now we are t+2_weeks?
This depends on your definition of noise. As long as the posts are CentOS related, I wouldn't care if there was 100 or 1000 posts per day.
1000 posts/day would certainly see me out. You're an obvious candidate to remain on all lists. I'm battling with my modem's load right now, I'm looking to cull what's less important.
John Summerfield wrote:
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I totally agree John - as CentOS gets more popular i too have found myself using 'Mark folder as Read' too often.
This is a good thing. I've just left the Fedora Core mailing lists (users & devel) and the traffic didn't bother me at all. It's a great
Not everyone can have such a fine Internet connexion.
There's always gmail if you'd rather have a web interface and not worry about the bulk of messages you skip.
Splitting by version is usually a bad idea. It causes things to split, people post the wrong version information to the wrong lists and it becomes a PITA. Hell, I've noticed enough people whinge and moan about top posting... Imagine that on someone posting a v5 question to a v4 list!
I thought it worked very well indeed with Red Hat Linux. I note that Red Hat continues with that plan with RHEL.
What happens if the person who knows the answer to a version 3 question has moved on to version 4? Or questions about the 90+% of things that are identical across all unix-like systems?
Les Mikesell wrote:
John Summerfield wrote:
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I totally agree John - as CentOS gets more popular i too have found myself using 'Mark folder as Read' too often.
This is a good thing. I've just left the Fedora Core mailing lists (users & devel) and the traffic didn't bother me at all. It's a great
Not everyone can have such a fine Internet connexion.
There's always gmail if you'd rather have a web interface and not worry about the bulk of messages you skip.
I can't read offline.
Splitting by version is usually a bad idea. It causes things to split, people post the wrong version information to the wrong lists and it becomes a PITA. Hell, I've noticed enough people whinge and moan about top posting... Imagine that on someone posting a v5 question to a v4 list!
I thought it worked very well indeed with Red Hat Linux. I note that Red Hat continues with that plan with RHEL.
What happens if the person who knows the answer to a version 3 question has moved on to version 4? Or questions about the 90+% of things that are identical across all unix-like systems?
Nothing's perfect. Why not everyone join the Fedora list? Just about everything that's in CentOS has been in Fedora too.
As I said, it seems to work in other places.
Ioannis Vranos wrote:
Les Mikesell wrote:
There's always gmail if you'd rather have a web interface and not worry about the bulk of messages you skip.
That's a great suggestion. Web email sounds as a great solution to John's problems.
Web browsing sucks. It often takes me minutes to download pages. Reread my ping times (those were good, here's another set) and how much I download.
[summer@bilby ~]$ ping -c4 terad.net PING terad.net (203.15.140.104) 56(84) bytes of data. 64 bytes from terad.net (203.15.140.104): icmp_seq=0 ttl=247 time=777 ms 64 bytes from terad.net (203.15.140.104): icmp_seq=1 ttl=247 time=2865 ms
--- terad.net ping statistics --- 4 packets transmitted, 2 received, 50% packet loss, time 11627ms rtt min/avg/max/mdev = 777.745/1821.475/2865.206/1043.731 ms, pipe 4 [summer@bilby ~]$
My mail arrives and I can read at LAN speeds.
John Summerfield wrote:
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
We've been down this route before - and the same things are going to be said now that were said previously, we are not going to be splitting the list based on Release and/or Arch. It might have worked for others, but we dont want to split the community up into fragments. And, personally, I think this is working very well - we get max eyeballs and people dont need to be subscribed to a multiple lists to keep an eye on stuff.
Perhaps when membership of the list hits 50,000 we might reconsider - but were not there yet.
What might be worth looking at was to create more lists ( not split ) based on technology (eg. CentOS-Clustering ), but we've not really had any request for those, afict.
And as Les has already pointed out - most stuff in the lists is common.
perhaps you should consider compress'd digests for your mailing lists traffic.
- KB
Karanbir Singh wrote:
John Summerfield wrote:
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
We've been down this route before - and the same things are going to be said now that were said previously, we are not going to be splitting the list based on Release and/or Arch. It might have worked for others, but we dont want to split the community up into fragments. And, personally, I think this is working very well - we get max eyeballs and people dont need to be subscribed to a multiple lists to keep an eye on stuff.
Perhaps when membership of the list hits 50,000 we might reconsider - but were not there yet.
What might be worth looking at was to create more lists ( not split ) based on technology (eg. CentOS-Clustering ), but we've not really had any request for those, afict.
And as Les has already pointed out - most stuff in the lists is common.
perhaps you should consider compress'd digests for your mailing lists traffic.
Okay, now we have a decision I'm outta here.
You said I'm not the first to raise the matter. You need to listen and consider the disadvantaged amongst your number.
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
You said I'm not the first to raise the matter. You need to listen and consider the disadvantaged amongst your number.
True. But that number is very small compared to the rest of the population. I would (no offense to you) rather sacrifice a few users who are still using (forced or otherwise) caveman technology to connect to the internet than fragment the knowledgebase and require folks to join a dozen or so mailing lists. If/when the mailing list reaches critical mass and a split is required for sanity, then so be it. I don't think we're there yet, and I completely agree with Karanbir on this one.
Jim Perrin wrote:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
You said I'm not the first to raise the matter. You need to listen and consider the disadvantaged amongst your number.
True. But that number is very small compared to the rest of the population. I would (no offense to you) rather sacrifice a few users
If you don't want to cause offence, the don't use pejorative terms.
who are still using (forced or otherwise) caveman technology to connect to the internet than fragment the knowledgebase and require folks to join a dozen or so mailing lists. If/when the mailing list reaches critical mass and a split is required for sanity, then so be it. I don't think we're there yet, and I completely agree with Karanbir on this one.
A problem is that the "voters," to some extent are self-selecting. Those who can't or won't deal with it leave and their opinions, if they voiced them at all, are forgotten.
This won't be the first list I leave because of the volume of email, but it's the one I cared about most.
John Summerfield wrote:
A problem is that the "voters," to some extent are self-selecting. Those who can't or won't deal with it leave and their opinions, if they voiced them at all, are forgotten.
I think the issue that there might be a few people ( and to be honest there have only been very few people in the last 3 years to say this ) who want a split by Release/Arch is outweighed by the fact that a single list is able to provide the best knowledge pool in the community at this time. Just creating more lists only fragments that up.
Also, if I may make a slightly personal statement - you seem to respond to most emails, regardless of what version or arch it might be targeting, so I take it that even if the lists were being split you might join all of them !
- KB
Karanbir Singh wrote:
John Summerfield wrote:
Also, if I may make a slightly personal statement - you seem to respond to most emails, regardless of what version or arch it might be targeting, so I take it that even if the lists were being split you might join all of them !
If I were, it would be my choice, and I would have that control.
I like to try to help, and I try to maintain good humour.
I don't see a reason to split on arch, the alternative architecture I know best (and my experience there is very old, around thirty years) is IBM's mainframes. Those folk have their own list, and any CentOS user running zSeries should be asking questions there (and they do), because their environment is so different - disk transfers at gigabytes/sec, they tend to run hundreds of penguins on one real system, their CPUs are relatively low-performance, mostly they're using virtual disks (they say "DASD"), they have virtual LANs and they have enterprise-grade policies and procedures to deal with.
I don't know where iSeries and pSeries users go, but I've not noticed any here (or on Nahant-list). I don't think I would expect them, their hardware environment is very different from IA32 and AMD-64 systems.
there's also digest mode..as was previously mentioned as well as the web-based archives.
John Summerfield wrote:
Jim Perrin wrote:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
You said I'm not the first to raise the matter. You need to listen and consider the disadvantaged amongst your number.
True. But that number is very small compared to the rest of the population. I would (no offense to you) rather sacrifice a few users
If you don't want to cause offence, the don't use pejorative terms.
who are still using (forced or otherwise) caveman technology to connect to the internet than fragment the knowledgebase and require folks to join a dozen or so mailing lists. If/when the mailing list reaches critical mass and a split is required for sanity, then so be it. I don't think we're there yet, and I completely agree with Karanbir on this one.
A problem is that the "voters," to some extent are self-selecting. Those who can't or won't deal with it leave and their opinions, if they voiced them at all, are forgotten.
This won't be the first list I leave because of the volume of email, but it's the one I cared about most.
Jim Perrin wrote:
True. But that number is very small compared to the rest of the population. I would (no offense to you) rather sacrifice a few users who are still using (forced or otherwise) caveman technology to connect to the internet than fragment the knowledgebase and require folks to join a dozen or so mailing lists. If/when the mailing list reaches critical mass and a split is required for sanity, then so be it. I don't think we're there yet, and I completely agree with Karanbir on this one.
However I think we should try to be accessible to the disadvantaged, GNU/Linux is about them too. Compressed digests sound as an option, given that text can be compressed very much. Is there any way to use compressed digests? I may even try using it myself.
John Summerfield wrote:
Think of this as a motion as one might move at a meeting. Discussion and refinement are in order.
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I am not on the Fedora users' list because of the volume of email, though I do use Fedora Core and might usefully contribute to FC in that way.
Similarly, I'm no longer on the OpenSUSE list, same story.
The time has come when I must, again, reduce the volume of email I see, because it's clogging my modem.
I use CentOS4 and plan to use CentOS5. I am interested in email for those, but not for CentOS 3 (I have no systems) nor CentOS 2 (I have one RHL 7.3 system that thinks it's CentOS 2, but it's in maintenance mode, and anything I do I will do alone).
I don't know how much email I would eliminate by dropping email for Centos < C4, but when I move to CentOS 5 I will lose interest in C4 and then there will be savings.
If this list were split into one for each release, then subscribers could choose which email they see. At present, it's all or none, and neither suits me.
If this list is split into four, then I expect the transition method would be to subscribe everyone on this list to the new four.
At some point, this would become read-only (nobody posts), or maybe all mail for this goes (via a filter to fix the headers) to all the others.
By "fix the headers" I mean "do whatever it takes to ensure replies go to the list the subscriber is replying to."
I'm sure this transition arrangement is imperfect; my objectives are to encourage CentOS to have the separate lists, and to ensure that the transition is fairly easy for users so we don't lose lots of subscribers.
What do others think?
For the time I have been in this list, I do not remember seeing any message about CentOS 3.x or older. I think what should be done is reducing the amount of emails per digest, so as to be able to use mime digests, which I think currently is not feasible, due to the amount of the messages per digest.
This is feasible, does not modify the size of emails, and is a correction to a mailing list "bug" (I consider not being able to use mime digests which is available as an option, a bug).
Ioannis Vranos wrote:
For the time I have been in this list, I do not remember seeing any message about CentOS 3.x or older.
So no need to split these off from this list :)
Although I remember answering a 2.1 question just a few weeks ago.
I think what should be done is reducing the amount of emails per digest, so as to be able to use mime digests, which I think currently is not feasible, due to the amount of the messages per digest.
It cannot be set by "number of mails before a digest is sent out", but only as "size in kB a digest should grow to before it is sent out".
This size is set fairly huge, so it never seems to be reached - 30 MB. That's why you're only seeing daily digests. I think this number could be set much lower if digest subscribers want to have that.
So other digest subscribers: Speak up!
This is feasible, does not modify the size of emails, and is a correction to a mailing list "bug" (I consider not being able to use mime digests which is available as an option, a bug).
Yes. But it might be a break in expected behaviour for people who only want one digest a day.
Cheers,
Ralph
Ralph Angenendt wrote:
to the amount of the messages per digest.
It cannot be set by "number of mails before a digest is sent out", but only as "size in kB a digest should grow to before it is sent out".
This size is set fairly huge, so it never seems to be reached - 30 MB.
Strewth! The default Postfix mailbox size is 10 Mbytes.
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
Jim Perrin spake the following on 5/1/2007 4:14 PM:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
Scott Silva wrote:
Jim Perrin spake the following on 5/1/2007 4:14 PM:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
A very bad joke, and a very sort point, but for me it's an unavoidable fact of life, at least until I move house.
John Summerfield spake the following on 5/1/2007 4:58 PM:
Scott Silva wrote:
Jim Perrin spake the following on 5/1/2007 4:14 PM:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
A very bad joke, and a very sort point, but for me it's an unavoidable fact of life, at least until I move house.
Sorry, I just couldn't resist. Any possibility of bonding two modem connections to get a little more bandwidth? Don't know if your ISP would allow it, but might be worth asking.
Scott Silva wrote:
Jim Perrin spake the following on 5/1/2007 4:14 PM:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
There are still people in Northern New Jersey (less than 30 miles from New York City), who still have no choice but to use a dial-up modem. I was talking to one of them last night.
Bisbal, Prentice wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
There are still people in Northern New Jersey (less than 30 miles from New York City), who still have no choice but to use a dial-up modem. I was talking to one of them last night.
If you are near a major highway you should be able to get a cell carrier's digital card, although the 'unlimited' data plans tend to be expensive and not really unlimited. I've found a few places my treo can't pick up email, but not many.
And there are always the satellite services like Starband if you are really out in the wilds - at least in North America.
On Wed, May 02, 2007 at 09:47:03AM -0400, Bisbal, Prentice said:
Scott Silva wrote:
Jim Perrin spake the following on 5/1/2007 4:14 PM:
On 5/1/07, John Summerfield Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
There are still people in Northern New Jersey (less than 30 miles from New York City), who still have no choice but to use a dial-up modem. I was talking to one of them last night.
Up until 2002, I lived in San Jose CA (Silicon Valley) and due to the distance to the CO, only had the option of modem (no 56K due to line quality - I usually got about 23kb) or IDSL which is 144Kb/144Kb for $140 / month + $400 install. I had the IDSL, but work paid for it :-) No cable modem service either.
You don't have to be rural to be unable to get affordable high-speed internet.
Bisbal, Prentice wrote:
Scott Silva wrote:
Jim Perrin spake the following on 5/1/2007 4:14 PM:
On 5/1/07, John Summerfield debian@herakles.homelinux.org wrote:
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
Anyone still using a modem is indeed a poor sod The rest of the intarweb doesn't seem to care about modem folk anymore either. We're just doing what all the other kids do :-P
What's a modem? ;-P
There are still people in Northern New Jersey (less than 30 miles from New York City), who still have no choice but to use a dial-up modem. I was talking to one of them last night.
We live about thirty miles West of Norfolk, VA and all that's available is dial-up. The cable company wanted $6000 to provide service, a matter of a few hundred yards/meters. Our only option is satellite service which works better than dial-up but not as good as the cable company service I had in Florida.
Bob Goodwin Zuni, Virginia
John Summerfield wrote:
Ralph Angenendt wrote:
It cannot be set by "number of mails before a digest is sent out", but only as "size in kB a digest should grow to before it is sent out".
This size is set fairly huge, so it never seems to be reached - 30 MB.
Strewth! The default Postfix mailbox size is 10 Mbytes.
50M.
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
That number is only for splitting off digests if that number is reached before the *daily* archive gets out. As I already said: This number can be changed as it really seems to be set too high.
But I want to hear from people if they would rather get *one* digest per day or up to three or four digests per day.
Ralph
Ralph Angenendt wrote:
That number is only for splitting off digests if that number is reached before the *daily* archive gets out. As I already said: This number can be changed as it really seems to be set too high.
But I want to hear from people if they would rather get *one* digest per day or up to three or four digests per day.
I think the digest as it is, it can not be used at all in the convenient MIME digest setting, that allows to reply easily to individual messages inside the digest.
So my vote is towards decreasing the size of the digest, as much as it is needed to use MIME digests easily.
--- Ioannis Vranos ivranos@freemail.gr wrote:
Ralph Angenendt wrote:
That number is only for splitting off digests if that number is reached before the *daily* archive gets out. As I already said: This number can be changed as it really seems to be set too high.
But I want to hear from people if they would rather get *one* digest per day or up to three or four digests per day.
I think the digest as it is, it can not be used at all in the convenient MIME digest setting, that allows to reply easily to individual messages inside the digest.
So my vote is towards decreasing the size of the digest, as much as it is needed to use MIME digests easily.
My vote is for 1 per day.
Ralph Angenendt spake the following on 5/2/2007 2:36 AM:
John Summerfield wrote:
Ralph Angenendt wrote:
It cannot be set by "number of mails before a digest is sent out", but only as "size in kB a digest should grow to before it is sent out".
This size is set fairly huge, so it never seems to be reached - 30 MB.
Strewth! The default Postfix mailbox size is 10 Mbytes.
50M.
Calculate how long it takes a 30 Mbyte message to go through a modem. Imagine if the poor sod waiting for it is downloading something else at the time.
That number is only for splitting off digests if that number is reached before the *daily* archive gets out. As I already said: This number can be changed as it really seems to be set too high.
But I want to hear from people if they would rather get *one* digest per day or up to three or four digests per day.
Ralph
I read through Gmane, because I never found a mail reader that worked with the mime digests either. I just let the newsreader connect and get the headers, and I only have to download the messages I really want to read.
Ioannis Vranos wrote:
John Summerfield wrote:
Think of this as a motion as one might move at a meeting. Discussion and refinement are in order.
I don't understand the current rationale for a single CentOS users' list; probably in times past it was sensible, but I think the time has come for splitting the list by release.
I'm speaking from my own perspective, but I'm sure others have similar stories. How many users use all CentOS releases?
I am not on the Fedora users' list because of the volume of email, though I do use Fedora Core and might usefully contribute to FC in that way.
Similarly, I'm no longer on the OpenSUSE list, same story.
The time has come when I must, again, reduce the volume of email I see, because it's clogging my modem.
I use CentOS4 and plan to use CentOS5. I am interested in email for those, but not for CentOS 3 (I have no systems) nor CentOS 2 (I have one RHL 7.3 system that thinks it's CentOS 2, but it's in maintenance mode, and anything I do I will do alone).
I don't know how much email I would eliminate by dropping email for Centos < C4, but when I move to CentOS 5 I will lose interest in C4 and then there will be savings.
If this list were split into one for each release, then subscribers could choose which email they see. At present, it's all or none, and neither suits me.
If this list is split into four, then I expect the transition method would be to subscribe everyone on this list to the new four.
At some point, this would become read-only (nobody posts), or maybe all mail for this goes (via a filter to fix the headers) to all the others.
By "fix the headers" I mean "do whatever it takes to ensure replies go to the list the subscriber is replying to."
I'm sure this transition arrangement is imperfect; my objectives are to encourage CentOS to have the separate lists, and to ensure that the transition is fairly easy for users so we don't lose lots of subscribers.
What do others think?
For the time I have been in this list, I do not remember seeing any message about CentOS 3.x or older. I think what should be done is
I have.
reducing the amount of emails per digest, so as to be able to use mime digests, which I think currently is not feasible, due to the amount of the messages per digest.
This is feasible, does not modify the size of emails, and is a correction to a mailing list "bug" (I consider not being able to use mime digests which is available as an option, a bug).
Digests don't, so far as I can estimate, help more than marginally, and on various lists I often see people reply to message "digest ...." I filter and <plonk> those.
The bulk of the messages is the same, there's just some variation in the volume of the headers.
-
Bisbal, Prentice -
Bob Goodwin - W2BOD -
BRUCE STANLEY -
Feizhou -
Ioannis Vranos -
Jim Perrin -
John R Pierce -
John Summerfield -
Karanbir Singh -
Kenneth Porter -
Les Mikesell -
MrKiwi -
Ralph Angenendt -
Scott Silva -
Steven Haigh -
Walt Reed -
William Warren