But what about bandwidth limits? John Companies seems to have a limit of 40Gb a month, whereas SF.net has no limit, at least none they list in their documentation.
Use a decent version-control system that can run over SSH. Create a repository on a central server, and "cvs checkout" onto every machine.
Then pick the files you want to share (bookmarks, bashrc, cron jobs, scripts, your TODO list, ~/bin), move them to the "cvs checkout"'d directory and leave symlinks to them from the old locations.
When you change them, do "cvs commit", and "cvs update" periodically on all machines.
Hey presto -- works great... occasionally you may have to fix conflicts by hand after a "cvs update", but otherwise it's brilliant.
Worth noting that we in the SA project will probably not be able to use these, as they've all been munged to remove some really excellent spam-signs: no Received headers, header case munged, etc.
I've been working on a similar project but using additional factors that help identify spam such as violations of the mail RFC's, and other header indicators, in addition to NLP. I have a prototype that I'm using to score all of my inbox e-mail and am using that to tune the weight factors and add in new factors as I encounter them. It would be interesting to combine your approach with mine I think, since I hadn't thought of analyzing trigrams.
Sounds a bit like SpamAssassin,
if I say so myself;)
SA analyses mail headers, body, and uses RBL and Razor to come up with an aggregate spam/non-spam
score, then filters appropriately. Most of its smarts is encapsulated in a Perl module, which
means it can be run from virtually anywhere; a procmail filter, a spam-protection SMTP proxy server, a system-wide checking system, etc. (all 3 of those have been implemented).
Its scores are generated using a GA and a large corpus of test mail, too. Hit rates nowadays are fantastic;)
I agree, but disagree at the same time. Somebody needs to be interested in the project, for it to live on. But if the project turns out to be
useful, they will make their own patches... and before you know it, someone's decided to merge the
patches together -- and hey presto, they're the "new maintainer". viz. Apache.
Anyway, regarding the initial Q.
My suggestion is this: ask around in the user
community, and see if anyone's interested
in taking over as a maintainer. If not,
forget about it -- just leave it
to gather dust. Put up a web page with the
current sources, details, etc. and a note
indicating that you no longer have time to
maintain it, and let it be.
If it's useful to other people, and it
works, and some tweak needs to be made,
somebody will make it. They may
also wind becoming de-facto maintainer
of the software (I know about this -- it
happened to me;)
If the new maintainer happens to irritate
its users, the source is still out there,
and if it has a decent free license, they
can pick their own new maintainer
and fork it. Maybe at some point down
the line they'll merge back in (egcs/gcc), maybe not (NCSA httpd/apache) -- it doesn't matter
as long as the code is good.
All of this is well and good -- it's the open source project lifecycle! Don't worry about it!
If your project is useful, it will
be used!
I know what I'm talking about here, because
I found a big, unmaintained project that
'scratched an itch' a time
ago;
I found the existing third-party patches out
there, integrated them all, made it portable
to new OSes, and released a new version.
Hey presto, I was the new (de-facto) maintainer.
It worked great. That's open source for ya.
Also, babbage said: This first guy said outright that a lot of people have downloaded his application but few have submitted patches back to it.
Depends on the project. Sysadmin/developer users generally submit patches, end-users don't. Also, if there's an active developer community, there's less incentive to 'scratch your own itch' when you can just throw in a suggestion and get it scratched for free;)
This is another key point -- while you're still
actively developing it, and appearing to "own" it, it will not pick up a new maintainer. You need to give people a need for
a maintainer, before one will appear!
I have previously seen a very expensive keyboard
designed to protect against RSI. It was similar
to this, in that the keys are split up and vertically mounted, but it was curved so that
the 2 keyboards were effectively on either side
of a cylinder.
Quite effective apparently, but v expensive.
I'm not sure about this one though. As a previous poster noted, it would be more RSI-friendly if it was curved.
There's another time ramdisks are good BTW -- when you're running from flash memory or a read-only HD or CD-ROM. In that situation, ramdisks have to be used for/tmp, and anywhere else you'll be storing data.
Sure, you can save data to flash filesystems, but flash memory has a limited number of writes -- so you cache as much as possible to/tmp or RAM, and only write to the flash when you're really really sure you want to do that.
I'm with John on this one -- fancy graphics engines do nothing but wave the developers' willies about. The Quake series of games were absolute turds for single-player gameplay (although I must admit I do enjoy a good multiplayer deathmatch;)
Anyway, for proof that a good gfx engine is not required for gameplay, look at Nethack -- an @ sign roaming a screenful of ASCII art. Yet the game is incredibly addictive, purely because you can do almost anything you like -- every zap a polymorph wand off a wall and turned yourself into an umber hulk, then chomped your way through solid rock? Now that's flexibility!
One thing I'd say about NH is that it needs more story. Some of the other roguelike games have taken care of that though. But IMHO the pure flexibility of the game makes up for it.
However, quite a few of us have already ditched AvantGo in favour of Sitescooper, which is a web-clipping engine oriented towards packaging up news sites for view offline on a PDA, using a DOC or HTML-based viewer. It's a lot smarter and more configurable than AG, and it's GPLed.
Among the hundreds of sites supported is Slashdot, which is displayed showing all stories with a comment filter at level 3. Check it out. If you have a Palm with a DOC reader or iSilo installed, you can pick up the previous day's news scooped into iSilo and DOC formats.
(Disclaimer: I'm the main author of sitescooper, and myself and Kornelis Sietsma came up with the site file for slashdot.)
XFree86 4, in CVS, includes code to do X with a resulting binary of about 600Kb, apparently. That sound pretty anti-bloat to me.
Personally I don't see a need to take X behind the shed and shoot it -- it has a good extension system to provide new extensions like anti-aliasing, alpha-blending, etc. -- and there are people working on the opposite direction, stripping it down.
And the network transparency is an incredible feature.
That chimes with my experiences with my Palm Vx -- I had to get a replacement cradle shipped. The previous Vx had a faulty touchscreen, so that had to be replaced, and now this.
This problem matches some reset problems I've had too, so I'll have to see if my Palm has this faulty DRAM.
I disagree. It would be perfectly feasible to build a webpad now, it'd just cost a fortune.
Basically the webpad reference designs and all the clones thereof require:
a CPU, RAM and flash that can support a decent web browser (HTML 3.2, frames, Javascript? SSL?). Flash memory at the moment is apparently very pricey.
a reasonably-sized color LCD screen
rechargeable battery power
short-range networking using the DECT cordless-phone technology, at the moment
My guess would be that to build this hardware at the moment, you'd wind up with an >$1000 price tag, once you include software costs, tax and a profit margin.
Also, another factor is that the Linux HTML4 browsers don't exist in a form factor that'll fit on embedded devices -- don't forget, these things don't have swap space at all. How big's your netscape running at the moment? Mine's at 35 megs. Ouch. And that's NS4.7, you should see what Mozilla does!;)
Having said all that... maybe there is a market for it in the T3 readership... but would it provide a big enough market to keep a company running?
Slashdot Mirror
But what about bandwidth limits? John Companies seems to have a limit of 40Gb a month, whereas SF.net has no limit, at least none they list in their documentation.
Use a decent version-control system that can run over SSH. Create a repository on a central server, and "cvs checkout" onto every machine.
Then pick the files you want to share (bookmarks, bashrc, cron jobs, scripts, your TODO list, ~/bin), move them to the "cvs checkout"'d directory and leave symlinks to them from the old locations.
When you change them, do "cvs commit", and "cvs update" periodically on all machines.
Hey presto -- works great... occasionally you may have to fix conflicts by hand after a "cvs update", but otherwise it's brilliant.
Worth noting that we in the SA project will probably not be able to use these, as they've all been munged to remove some really excellent spam-signs: no Received headers, header case munged, etc.
Sounds a bit like SpamAssassin, if I say so myself ;)
SA analyses mail headers, body, and uses RBL and Razor to come up with an aggregate spam/non-spam score, then filters appropriately. Most of its smarts is encapsulated in a Perl module, which means it can be run from virtually anywhere; a procmail filter, a spam-protection SMTP proxy server, a system-wide checking system, etc. (all 3 of those have been implemented). Its scores are generated using a GA and a large corpus of test mail, too. Hit rates nowadays are fantastic ;)
Disclaimer: I'm the maintainer.
Anyway, regarding the initial Q. My suggestion is this: ask around in the user community, and see if anyone's interested in taking over as a maintainer. If not, forget about it -- just leave it to gather dust. Put up a web page with the current sources, details, etc. and a note indicating that you no longer have time to maintain it, and let it be.
If it's useful to other people, and it works, and some tweak needs to be made, somebody will make it. They may also wind becoming de-facto maintainer of the software (I know about this -- it happened to me ;)
If the new maintainer happens to irritate its users, the source is still out there, and if it has a decent free license, they can pick their own new maintainer and fork it. Maybe at some point down the line they'll merge back in (egcs/gcc), maybe not (NCSA httpd/apache) -- it doesn't matter as long as the code is good.
All of this is well and good -- it's the open source project lifecycle! Don't worry about it! If your project is useful, it will be used!
I know what I'm talking about here, because I found a big, unmaintained project that 'scratched an itch' a time ago; I found the existing third-party patches out there, integrated them all, made it portable to new OSes, and released a new version. Hey presto, I was the new (de-facto) maintainer. It worked great. That's open source for ya.
Also, babbage said: This first guy said outright that a lot of people have downloaded his application but few have submitted patches back to it.
Depends on the project. Sysadmin/developer users generally submit patches, end-users don't. Also, if there's an active developer community, there's less incentive to 'scratch your own itch' when you can just throw in a suggestion and get it scratched for free ;)
This is another key point -- while you're still actively developing it, and appearing to "own" it, it will not pick up a new maintainer. You need to give people a need for a maintainer, before one will appear!
The patent's claims, as described, are a perfect description of a web browser; all browsers, from Mosaic on, have been built using those components.
Quite effective apparently, but v expensive.
I'm not sure about this one though. As a previous poster noted, it would be more RSI-friendly if it was curved.
I've converted the book into a HTML version, and from there into Plucker and iSilo (two e-book formats for the Palm). Pick 'em up here!
Agreed here too.
/tmp, and anywhere else you'll be storing data.
/tmp or RAM, and only write to the flash when you're really really sure you want to do that.
There's another time ramdisks are good BTW -- when you're running from flash memory or a read-only HD or CD-ROM. In that situation, ramdisks have to be used for
Sure, you can save data to flash filesystems, but flash memory has a limited number of writes -- so you cache as much as possible to
absolute turds for single-player gameplay (although I must admit I do enjoy a good multiplayer deathmatch
Anyway, for proof that a good gfx engine is not required for gameplay, look at Nethack -- an @ sign roaming a screenful of ASCII art. Yet the game is incredibly addictive, purely because you can do almost anything you like -- every zap a polymorph wand off a wall and turned yourself into an umber hulk, then chomped your way through solid rock? Now that's flexibility!
One thing I'd say about NH is that it needs more story. Some of the other roguelike games have taken care of that though. But IMHO the pure flexibility of the game makes up for it.
However, quite a few of us have already ditched AvantGo in favour of Sitescooper, which is a web-clipping engine oriented towards packaging up news sites for view offline on a PDA, using a DOC or HTML-based viewer. It's a lot smarter and more configurable than AG, and it's GPLed.
Among the hundreds of sites supported is Slashdot, which is displayed showing all stories with a comment filter at level 3. Check it out. If you have a Palm with a DOC reader or iSilo installed, you can pick up the previous day's news scooped into iSilo and DOC formats.
(Disclaimer: I'm the main author of sitescooper, and myself and Kornelis Sietsma came up with the site file for slashdot.)
XFree86 4, in CVS, includes code to do X with a resulting binary of about 600Kb, apparently. That sound pretty anti-bloat to me.
Personally I don't see a need to take X behind the shed and shoot it -- it has a good extension system to provide new extensions like anti-aliasing, alpha-blending, etc. -- and there are people working on the opposite direction, stripping it down.
And the network transparency is an incredible feature.
That chimes with my experiences with my Palm Vx -- I had to get a replacement cradle shipped. The previous Vx had a faulty touchscreen, so that had to be replaced, and now this.
This problem matches some reset problems I've had too, so I'll have to see if my Palm has this faulty DRAM.
Definite quality issues there...
Basically the webpad reference designs and all the clones thereof require:
- a CPU, RAM and flash that can support a decent web browser (HTML 3.2, frames, Javascript? SSL?). Flash memory at the moment is apparently very pricey.
- a reasonably-sized color LCD screen
- rechargeable battery power
- short-range networking using the DECT cordless-phone technology, at the moment
My guess would be that to build this hardware at the moment, you'd wind up with an >$1000 price tag, once you include software costs, tax and a profit margin.Also, another factor is that the Linux HTML4 browsers don't exist in a form factor that'll fit on embedded devices -- don't forget, these things don't have swap space at all. How big's your netscape running at the moment? Mine's at 35 megs. Ouch. And that's NS4.7, you should see what Mozilla does! ;)
Having said all that... maybe there is a market for it in the T3 readership... but would it provide a big enough market to keep a company running?
(BTW this is all my opinion, not employer stuff)