LOL. At first I thought you were making a Real Genius reference, but then I realized that his name was Laslow and I remembered my Psychology courses....
-matthew
Re:Perl Out, Ruby In - Thank God
on
The Ruby Way
·
· Score: 1
*shrug* Whatever. I have done Ruby work outside of Rails and I really have not had any problems with documentation except for specific libraries (usually beta stage). Maybe the CGI libraries are a special case? I don't think the documenation situation is nearly as bas as you suggest. But I guess I'm coming from the Linux world where documentation is generally lacking. Perhaps I am just used to diving into code now and then.
"The D alleles may not even change brain size; they may only make the brain a bit more efficient if it indeed affects brain function," Lahn said.
Err, so I guess this doesn't necessarily have anything to do with brain size OR intelligence. The only really significant result of the finding so far is that Homo Sapien and Neaderthal may have interbread. We don't even know what this "brain gene" does.... just that some people have it. It could make people more prone to schizophrenia for all we know. That is, until somebody actually tests for measurable, statistically significant differences between the 30% with and 70% without...
I think once YARV is ready for general consumption, it'll take off like wildfire.
Good luck on that. As far as I can tell, YARV has been all but dead the last few months. I've had the svn repo checked out and have seen very little activity. Very discouraging. I've had to give up using Ruby for a couple projects because of serious performance issues compared to, say, Python. And that is too bad because I can't say I care much for Python.
-matthew
Re:Perl Out, Ruby In - Thank God
on
The Ruby Way
·
· Score: 2, Interesting
Ruby has a lot of work in the library department, and a HELL of a lot of work in the documentation department.
I never really understood the documentation complaint about Ruby (I use Rails, mostly). Most everything I need to know about the core libs is in http://www.ruby-doc.org/ and http://api.rubyonrails.com/ is nearly complete. It is no worse than the PHP online documentation. Actually, it is better. I find RDoc much faster to search (CTRL-F). And you can get the first edition of the book, Programming Ruby: The Pragmatic Programmer's Guide online at http://www.rubycentral.com/book/ to get more in depth knowledge of how Ruby works. The IRC channels are very active and helpful (I help lots of people there myself). There is a great mailing list/form at http://www.ruby-forum.com/forum/3 And there are a million Rails tutorial out there. I learned Ruby/Rails to a point where I could do useful things in like 5 days without any prior exposure with no hassle. So what is missing, exactly?
-matthew
Re:I Really Like Ruby and Rails
on
The Ruby Way
·
· Score: 2, Interesting
The good news is that he doesn't have an index.php [f1newstoday.com], unlike certain other sites [rubyonrails.org]...
Please note that rails is a framework intended for web applications, not web sites with small bits of dynamic content. So it isn't necessarily damning that the Rails website is written in PHP. Use the right tool for the job. For simple, dynamic content PHP works just fine where Rail would be overkill.
Aha? I have not looked at dspam for quite some time, but for the setup and scale I have described here, we don't need a dedicated database server, just a few extra attributes in ldap. If you don't have ldap, files on the local filesystem should work for most smallish setups, but you may of course want to put it in a database for other reasons still if you already have one anyway.
In LDAP? Are you serious? With DSPAM, dictionaries grew to be tens of megabytes per user. I wasn't even aware that LDAP was capable of indexing attributes that large. You can't query LDAP like you can SQL or even BDB. I imagine for each mail scan, you'd have to download the entire user's dictionary from the directory. That sounds HORRIBLY inefficient. Or is there some capability of LDAP that I am missing?
It's annoying as all hell, and I wish I didn't have to put up with it - as it means the machine (which I donated to host a number of community wireless networking sites) will need an upgrade just to help with spam scanning. Something I really can't afford to do - not everyone can blow $2-3k on a new server to keep up with spam:(
Upgrade to scan 1000 messages a day? No way. My gateway scans nearly 4000 per day. You can see some stats here: http://mailgw.pnca.edu/cgi-bin/mailgraph.cgi . And that is running on a pretty low end machine. One important thing is to reject mail to unknown users before they get scanned. That keeps dictionary attacks from killing your server. Anvil (Postfix) also helps.
Are you just using the stock SA rules? Check out http://www.rulesemporium.com/ or if you run FreeBSD, install the spamass-rules port. I have not found a trained spamassassin to be terribly effective. It only adds a point or two to the total score. (although you can increase the importance of the Bayes score, of course) Using a wide variety of rules is more effective overall than bothering with training in my experience. But I do employ some amount of training. I have a shared IMAP box where users can dump spam that gets through. I have a script that periodically downloads the contents of that folder and trains SA... It does a point here and there. Also, SA will automatically train itself with spam that scores over a certain threshold.
I don't have statistics for everyone here at our org, but I don't receive more than one spam every couple days. Sometimes I'll go a week without seeing one. Combine that will Adblock Plus in Firefox and the Internet is a pleasant place again.;-)
A bayes filter only works when you have enough good data to feed it with.
The bayes filter that I've used, DSPAM, didn't work as well when it was heavily weighted towards SPAM. Even the 80% spam I (would) get today would be too much. Having a whole separate spam trap would be overkill.
At any rate, spamassassin supports bayes as you probably know, making good use of it does not depend or conflict with whatever spamassassin rulesets you happen to use as long as you do add a few statements to your local.cf file, in other words, it is on top of, not instead of.
Unless by adding a few statements to local.cf, you mean add weight to the Bayes tests, I don't think the bayes support in SA is very useful. I don't remember the scoring off hand, but I think it is only like +1.5 even for 99% certainty. Enough to push a few edge cases over the threshold, but not very significant when most spam scores 10+ from all the other rules. Also, it is a server wide/shared bayes dictionary (at least when set up with amavisd). Group dictionaries are not nearly as effective as personal dictionaries. To have a truely effective bayes filter, you need to maintain a dictionary for each and every user. And THAT is some pretty big resources. When I was using DSPAM, i had to dedicate a database server just for the user dictionaries. Big pain in the ass.
How can we disagree? I don't even think we are having the same conversation. It might help if you actually responded to my points directly instead of pretending they never existed. You seem to think that I am expounding upon the virtues of Diablo. I'm just saying that it isn't in the same class as roguelikes like Nethack or Angband. Not because rouguelikes are better or more interesting (although I agree that they are), but just because roguelikes are a class of their own. The styles of gameplay are just different. There are, in fact, graphical frontends for Nethack. And they aren't Diablo.
Reminds me of when I first installed SpamAssassin on my mail server:)
Of course today, no matter what I do, the majority still gets through.
Then your setup is broken. Works great here, even today. I did get a couple of the stock pump-n-dump scams a few days ago (possibly related to the botnet from the article), but a little tweaking took care of that.
By using such 'spam' accounts to trap spam and feed it to your spam filter for learning?
Still, why bother? I mean, unless you are developing SA rules or reporting to public blacklists. The default Spamassassin rules alone are pretty good. Add in SARE and some other public rules sets and you don't even need learning. I used to use bayesian learning but found that it was much more maintenance than it was worth. Quite the opposite of what I originally thought. I thought maintaining SA rules would be a pain, but it is actually simpler than managing huge dictionaries.
I never really understood why people go out of their way to create, delete, and otherwise hassle with "spam" accounts or dummy accounts when you can just have one address with good spam filtering. It just seems like a lot of unnecessary work. I run a Spamassassin gateway that catches nearly all SPAM (80% of all email is blocked). I don't have to worry about keeping my address secret. I use it all over the place. Forums, online transactions, and even Usenet. I see almost no spam. How could some convoluted account manipulation scheme be better?
I've been seeing over 80% SPAM in the last couple months. And that is just what is being blocked (spamassassin). The actual number is a little higher. Sad, really.
The problem with normal text mode is that you can't tell a ghost (" ") form a darkened portion of the room (also " ") without specifically looking around. Realistic, perhaps, but also annoying.
I think that is kinda the point.
I once, long ago, when i still used DOS, found a program that replaced the font with special symbols. It wasn't exactly graphical, but it was just enough to turn the game from ascetic to comfortable. Ah, the hours I spent chasing after succubi...
I agree, roguelikes are better with some sort of extended ASCII rendering (real line draw characters instead of the generic # walls, for example). The reason more roguelikes don't do it by default is portability. A lot of systems won't display extended ASCII the way DOS did.
Diablo started as a turn-based game, but the team made a (very good) decision to go real-time. I read an interview with the devs some years back.
Whether or not the developers started Diablo as turn-based is irrelevent. It isn't turn-based. The turn-based nature of Nethack (and since you brought it up, Andband) is one of the primary features that makes the (sub)genre we call "roguelike" in which Diablo is not included.
Diablo randomly generates a level full of monsters, including a handful of 'boss' monsters. Angband does the same, but the special levels are far harder and can contain many 'boss' monsters.
The *bands, and rouguelikes, in general are not nearly as focused on 'bosses' as Diablo was (note, I only played II). In roguelikes, they are often call "uniques." Nethack has surprisingly few uniques if you think about it. Rogue didn't have any that I am aware of. It is not a primary feature of a the (sub)genre. The focus in roguelikes is on the common monsters and trying not to die Yet Another Stupid Death. In Diablo, death is rather inconsequential. Just restore from a save game...
Diablo has a range of randomly-generated items scattered through the level and left behind by monsters. Angband has the same.
Methinks you are pushing this "randomly generated" commonality too far. It is common in nearly all RPG games for monsters to carry loot that is randomly generated to some degree or another.
Diablo has a very limited range of items and monsters. Angband has many times more.
Wow. They both have a range of items and monsters. Imagine that.
After thoroughly playing Diablo one and two, I got the feeling that it was "Angband lite" - prettier but with less substance. I sold my copies of D1 and D2 and returned to Angband. I was much happier, although I was forced to imagine more.
No, Rogue was Angband-lite (if it is fair to compare games a decade or more apart). Diablo was a real-time RPG that allowed you to save your progression.
Of course, the 3D interfaces are difficult to play. But maybe I'm must too used to having the full dumgeon map laid out in small, succinct characters for maximum information. Falcon's Eye, while nice looking, was unplayable, IMO.
Mind if I ask what you need a GUI for? If I remember correctly, the Windows version isn't significantly different than the old fashioned console version. Do you need tiles? Personally, I won't use anything that doesn't show monsters as letters. There is no better way to identify/classify the myriad of monsters in Nethack than letters. The tiles are usually too small to be very informative. And not knowing what you are up against is usually fatal.
You should at least be able to get to the castle (killed by a drawbridge lately?) often. Although, sadly, I often don't make it through the gnomish mines. Once I make it through there, i'm usually in for a good long game.
My favorite character is a valkyrie. Gauntlets of Power + Mjollnir == serious ass kicking.
Although I don't play Nethack anymore. I'm a total ToME convert.
I disagree. I'd hire good managers. Seriously. I know this won't be received well here on/., but it's true: Having a good manager (one who can get things done quickly and efficiently) is not a bad thing when it comes to the bottom line (profit).
So how exactly does this solve the technical and capacity shortcomings of their current IT staff. Yeah, good managers are important, but if you don't the staff, you can't take on bigger clients, PERIOD.
i agree, your staffing issues are your primary concern, without additional staff you are not going to be able to expand at all. Finding staff is hard, but as discussed in a recent slashdot article, the key is paying well. If there is competition for skilled people, the company who is paying the most will win.
This is not true in my case. My primary concern when looking for an IT job is finding an environment that fits my personality best. I am currently making nearly *half* of what I potentially could just to work in a smaller, more flexible, more laid back environment. Also, we just recently hired a good IT person who made a similar choice.
Smaller companies can get often away with not paying as much because they offer much more in other ways. But hey, if you like endless meetings, cubical politics, and sucking up to incompetent middle management, all for a bigger paycheck, go for it.
Take that Maslow.
LOL. At first I thought you were making a Real Genius reference, but then I realized that his name was Laslow and I remembered my Psychology courses....
-matthew
*shrug* Whatever. I have done Ruby work outside of Rails and I really have not had any problems with documentation except for specific libraries (usually beta stage). Maybe the CGI libraries are a special case? I don't think the documenation situation is nearly as bas as you suggest. But I guess I'm coming from the Linux world where documentation is generally lacking. Perhaps I am just used to diving into code now and then.
-matthew
Err, so I guess this doesn't necessarily have anything to do with brain size OR intelligence. The only really significant result of the finding so far is that Homo Sapien and Neaderthal may have interbread. We don't even know what this "brain gene" does.... just that some people have it. It could make people more prone to schizophrenia for all we know. That is, until somebody actually tests for measurable, statistically significant differences between the 30% with and 70% without...
-matthew
Good luck on that. As far as I can tell, YARV has been all but dead the last few months. I've had the svn repo checked out and have seen very little activity. Very discouraging. I've had to give up using Ruby for a couple projects because of serious performance issues compared to, say, Python. And that is too bad because I can't say I care much for Python.
-matthew
I never really understood the documentation complaint about Ruby (I use Rails, mostly). Most everything I need to know about the core libs is in http://www.ruby-doc.org/ and http://api.rubyonrails.com/ is nearly complete. It is no worse than the PHP online documentation. Actually, it is better. I find RDoc much faster to search (CTRL-F). And you can get the first edition of the book, Programming Ruby: The Pragmatic Programmer's Guide online at http://www.rubycentral.com/book/ to get more in depth knowledge of how Ruby works. The IRC channels are very active and helpful (I help lots of people there myself). There is a great mailing list/form at http://www.ruby-forum.com/forum/3 And there are a million Rails tutorial out there. I learned Ruby/Rails to a point where I could do useful things in like 5 days without any prior exposure with no hassle. So what is missing, exactly?
-matthew
Please note that rails is a framework intended for web applications, not web sites with small bits of dynamic content. So it isn't necessarily damning that the Rails website is written in PHP. Use the right tool for the job. For simple, dynamic content PHP works just fine where Rail would be overkill.
-matthew
In LDAP? Are you serious? With DSPAM, dictionaries grew to be tens of megabytes per user. I wasn't even aware that LDAP was capable of indexing attributes that large. You can't query LDAP like you can SQL or even BDB. I imagine for each mail scan, you'd have to download the entire user's dictionary from the directory. That sounds HORRIBLY inefficient. Or is there some capability of LDAP that I am missing?
-matthew
Upgrade to scan 1000 messages a day? No way. My gateway scans nearly 4000 per day. You can see some stats here: http://mailgw.pnca.edu/cgi-bin/mailgraph.cgi . And that is running on a pretty low end machine. One important thing is to reject mail to unknown users before they get scanned. That keeps dictionary attacks from killing your server. Anvil (Postfix) also helps.
-matthew
Are you just using the stock SA rules? Check out http://www.rulesemporium.com/ or if you run FreeBSD, install the spamass-rules port. I have not found a trained spamassassin to be terribly effective. It only adds a point or two to the total score. (although you can increase the importance of the Bayes score, of course) Using a wide variety of rules is more effective overall than bothering with training in my experience. But I do employ some amount of training. I have a shared IMAP box where users can dump spam that gets through. I have a script that periodically downloads the contents of that folder and trains SA... It does a point here and there. Also, SA will automatically train itself with spam that scores over a certain threshold.
;-)
I don't have statistics for everyone here at our org, but I don't receive more than one spam every couple days. Sometimes I'll go a week without seeing one. Combine that will Adblock Plus in Firefox and the Internet is a pleasant place again.
-matthew
The bayes filter that I've used, DSPAM, didn't work as well when it was heavily weighted towards SPAM. Even the 80% spam I (would) get today would be too much. Having a whole separate spam trap would be overkill.
Unless by adding a few statements to local.cf, you mean add weight to the Bayes tests, I don't think the bayes support in SA is very useful. I don't remember the scoring off hand, but I think it is only like +1.5 even for 99% certainty. Enough to push a few edge cases over the threshold, but not very significant when most spam scores 10+ from all the other rules. Also, it is a server wide/shared bayes dictionary (at least when set up with amavisd). Group dictionaries are not nearly as effective as personal dictionaries. To have a truely effective bayes filter, you need to maintain a dictionary for each and every user. And THAT is some pretty big resources. When I was using DSPAM, i had to dedicate a database server just for the user dictionaries. Big pain in the ass.
-matthew
How can we disagree? I don't even think we are having the same conversation. It might help if you actually responded to my points directly instead of pretending they never existed. You seem to think that I am expounding upon the virtues of Diablo. I'm just saying that it isn't in the same class as roguelikes like Nethack or Angband. Not because rouguelikes are better or more interesting (although I agree that they are), but just because roguelikes are a class of their own. The styles of gameplay are just different. There are, in fact, graphical frontends for Nethack. And they aren't Diablo.
-matthew
Then your setup is broken. Works great here, even today. I did get a couple of the stock pump-n-dump scams a few days ago (possibly related to the botnet from the article), but a little tweaking took care of that.
-matthew
Still, why bother? I mean, unless you are developing SA rules or reporting to public blacklists. The default Spamassassin rules alone are pretty good. Add in SARE and some other public rules sets and you don't even need learning. I used to use bayesian learning but found that it was much more maintenance than it was worth. Quite the opposite of what I originally thought. I thought maintaining SA rules would be a pain, but it is actually simpler than managing huge dictionaries.
-matthew
I never really understood why people go out of their way to create, delete, and otherwise hassle with "spam" accounts or dummy accounts when you can just have one address with good spam filtering. It just seems like a lot of unnecessary work. I run a Spamassassin gateway that catches nearly all SPAM (80% of all email is blocked). I don't have to worry about keeping my address secret. I use it all over the place. Forums, online transactions, and even Usenet. I see almost no spam. How could some convoluted account manipulation scheme be better?
-matthew
I've been seeing over 80% SPAM in the last couple months. And that is just what is being blocked (spamassassin). The actual number is a little higher. Sad, really.
-matthew
I think that is kinda the point.
I agree, roguelikes are better with some sort of extended ASCII rendering (real line draw characters instead of the generic # walls, for example). The reason more roguelikes don't do it by default is portability. A lot of systems won't display extended ASCII the way DOS did.
-matthew
Whether or not the developers started Diablo as turn-based is irrelevent. It isn't turn-based. The turn-based nature of Nethack (and since you brought it up, Andband) is one of the primary features that makes the (sub)genre we call "roguelike" in which Diablo is not included.
Lots of games randomly generate things.
The *bands, and rouguelikes, in general are not nearly as focused on 'bosses' as Diablo was (note, I only played II). In roguelikes, they are often call "uniques." Nethack has surprisingly few uniques if you think about it. Rogue didn't have any that I am aware of. It is not a primary feature of a the (sub)genre. The focus in roguelikes is on the common monsters and trying not to die Yet Another Stupid Death. In Diablo, death is rather inconsequential. Just restore from a save game...
Methinks you are pushing this "randomly generated" commonality too far. It is common in nearly all RPG games for monsters to carry loot that is randomly generated to some degree or another.
Wow. They both have a range of items and monsters. Imagine that.
No, Rogue was Angband-lite (if it is fair to compare games a decade or more apart). Diablo was a real-time RPG that allowed you to save your progression.
-matthew
Oh please. They're barely in the same genre. For one thing, Diablo is real-time/action where Nethack is turn based/strategy...
-matthew
Of course, the 3D interfaces are difficult to play. But maybe I'm must too used to having the full dumgeon map laid out in small, succinct characters for maximum information. Falcon's Eye, while nice looking, was unplayable, IMO.
-matthew
Mind if I ask what you need a GUI for? If I remember correctly, the Windows version isn't significantly different than the old fashioned console version. Do you need tiles? Personally, I won't use anything that doesn't show monsters as letters. There is no better way to identify/classify the myriad of monsters in Nethack than letters. The tiles are usually too small to be very informative. And not knowing what you are up against is usually fatal.
-matthew
You should at least be able to get to the castle (killed by a drawbridge lately?) often. Although, sadly, I often don't make it through the gnomish mines. Once I make it through there, i'm usually in for a good long game.
My favorite character is a valkyrie. Gauntlets of Power + Mjollnir == serious ass kicking.
Although I don't play Nethack anymore. I'm a total ToME convert.
-matthew
Even a smaller object experiences air friction. It still increases by V^2.
-matthew
Yeah, and if I had a billion dollars, I wouldn't need to work anymore.
-matthew
So how exactly does this solve the technical and capacity shortcomings of their current IT staff. Yeah, good managers are important, but if you don't the staff, you can't take on bigger clients, PERIOD.
-matthew
This is not true in my case. My primary concern when looking for an IT job is finding an environment that fits my personality best. I am currently making nearly *half* of what I potentially could just to work in a smaller, more flexible, more laid back environment. Also, we just recently hired a good IT person who made a similar choice.
Smaller companies can get often away with not paying as much because they offer much more in other ways. But hey, if you like endless meetings, cubical politics, and sucking up to incompetent middle management, all for a bigger paycheck, go for it.
-matthew