Usually I subscribe to the philosophy of "what goes on in slashdot stays in slashdot." But since we had an interesting discussion in the irc channel reguarding RAM and kernel configurables, I suppose I can use this as a jumping board for further discussion.
Frankly, OO.o suffers the same amount of salesmanship when presented by OSS advocates as any other piece of software. And a threatening blog post won't effectively change anything about it, even when given high visibility via slashdot. Partly, you need the will to change, not just the will to criticize. Bootchart.org is a good example. Someone suspected that Linux boot time was bad and put forth a challenge to the general public to cobble together a tool to graphically represent the boot process. The result was a tool easily adopted for various distributions, and across the board competition among distros to eliminate obvious bottlenecks. This cases feels related; both need some metrics and visualization before significant improvements can be made. In neither OO.o nor boot time has vocal dissatisfaction with Free software measurably improved things.
But on to the interesting bit. Given that we have copious amounts of RAM these days, and long uptimes, many people like to work around the problem of OO.o and other memory hogs by simply buying tons of memory and leaving the apps open at all times. They reason that with enough RAM VM shouldn't swap out their programs, even though they don't use them often. The frequent foil to this however, is the dreaded disk intensive cron job. A hungry updatedb process can easily chew through disk cache and scare an unused OO.o out of RAM in the quest for more cache. The result is a ponderous computer when you wake up in the morning or after any disk intensive activity.
One solution, that a fellow LUG'er found useful was the kernel's swappiness setting. This parameter affects a kernel's desire to swap things out. People who utterly despise swap might consider setting swappinnes to zero. The maintainer has previously suggested on lkml that he runs his desktop at 100. I've decided to try it out, with no serious consequences so far, but no really noticable benefits either.
jldugger
On 10/27/05, Leo Mauler [email protected] wrote:
Saw this on /.
http://slashdot.org/articles/05/10/27/1425232.shtml?tid=185
After spending an entire ITEC exposition extolling the virtues of OOo, this crops up to claim that OOo has memory usage issues and very slow startup times.
I'll admit that I haven't used M$ Office all that much, but I don't seem to notice a significant time period between running OOo and starting to write a document or create a spreadsheet.
Has anyone else who uses both noticed the kind of lag on OOo that the article author describes?
The Slashdot replies also go into the point that most people have M$ Office already, legally or illegally, so we can't talk "free" without running up against all the illegal copies out there which were also, technically, "free".
One Slashdot reply says that Longhorn will make the "free" case for OOo by preventing people from running illegal copies of M$ software. Does anyone know about this in more detail?
Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com _______________________________________________ Kclug mailing list [email protected] http://kclug.org/mailman/listinfo/kclug