Why do people criticise, say, checkpoint firewalls, for giving out the fact that they are checkpoint firewalls, when you telnet into them? The point is not just "ooh look, there's a port open" but rather that the information has been let loose, "so goodness knows what's going to happen". (There's a reason I didn't let on anything particularly interesting with my two URL links!)
So there's a firewall in the way (?solaris?iss?) and there are non-80 ports that are filtered, while 80 is the only one open...
Sure, maybe the worst that can happen is a remote DoS attack - quite possibly the case, but I'm not going to launch *that* - but do you really want the Royal webpages getting defaced and/or being unavailable, even for half an hour?
RedHat 5.2 is ancient, apache 1.3.3 is buggy (insecure? quite possibly)... so thanks for the information!
Just an idea... I wouldn't have let loose what it was running on, myself - but then again, it would be interesting to see if they have crack-attempt logs in the manner of antionline, and so on...
Not *everyone* sweats away listening to dodgy loud metal trying to get the latest patch to apache to work, or whatever they're hacking on, for free. Businesses still exist and like to make money - it's way too early for us all to go broke & hippy about it all. Admittedly, M$loth apps are the last things I want to run - I do, however, need NT at home for the scanner and quake2, but otherwise I work off linux all day - because my ideology is not fundamentalist "linux has to be better", but rather, it *is* the way I want to work.
I generally agree. The problem is "1 window manager, 1 desktop environment, 1 version" - we all know that with that limitation, as some would have it be, OpenSource goes out t'window. Open-Source is not linux, in some ways it's "more than" linux - it's the software development ethic, more than anything else. And FreeBSD (and others) are, as yet, unpolluted by the influx of marketroids.
Is it possible that one day, we'll all be saying "Open-Source is essential" and actually wake up & realise it doesn't happen any more, given the rate at which non-free things are being ported?
I've not done IE under unix (of any sort) recently, so can't update you on that.
But a comment on the browser scene: netscrape, sure, is underfeatured and over-bloated. IE is evil, of course. Mozilla isn't remotely ready for "public consumption". My recommendations on the unix browser scene boil doing to either Opera (as & when they get their act together) or Konqueror - part of the KDE. I'm using the latter now as one of my mainstays for non-javascript browsing - even slashdot comes out OK in it, and it's stable enough (1.1.2). Oh, and it also renders PNGs correctly in a webpage.
Yes indeed - so how come we have lots of different ways of slagging them off all marked up as "insightful" and various other things, and yet we're no good at praise where it's due? Sure it was the wrong thing to start with, but at least Corel listened to the community, or something, and have now rectified it. This is not bad - it's not necessarily as good as it could be, but it *does* show "something stirs", or whatever.
It's a misfeature of slashdot, IMO, that there is neither room for duplication of feeling nor for going that far off-topic.
Well, I don't know why everyone else on here is into heavy / metal / noisy type stuff (says me, choosing terminology so's not to offend:) but as for me, I like a nice bit of quiet RunRig or Wolfstone, for coding, newsgroups,...whatever.
Must be mad:)
Re:Desktop standards v. networking standards
on
Gartner Slams Linux
·
· Score: 1
Yeah, in that case, the desktop environment differences do cause some problems - at least 3 different window managers (enlightenment, kwm and wm) are popular with their assorted CORBA-using layers above (KDE, Gnome). However, it is quite possible to run several of these all at once - I've got KDE (with kwm) and a gnome panel sitting here quite happily. The fun & games only really start with gmc running as it's anyone's guess who gets the desktop clicks;)
As far as the average user approaching unix, linux, or X goes, there's bound to be some confusion. The term 'window manager' must be explained, at the very least. What we can't have is the sort of state of play, "what window manager are you using?" - "linux!" *duh*... And while I'm not exactly the best support / trainer around, if the 30-yr-old unix model with X and stuff can finally take off, all to the good.
Oh, as for star office, I have actually seen folks asking where it should be installed - normally/usr/local/so51/ or similar is suggested, with ensuing wars about/usr/local and/usr and/opt and all that stuff. So there are users out there who want to do The Right Thing(TM), fortunately. And these I have a chance of getting on favourably with...:)
"True: Linux may not have many standards, but you can standardise on a distrobution. Decide on what one you think is best for your operation then stick with it."
The problem is, that's not true. We have RFCs, just like everything else. Some stuff implemented on linux doesn't live up to the RFCs in question (dhcpd 2, for example, doesn't implement DHCPINFORM), but then again it also implements some stuff far better than mere standards-compliance (gnutar, bzip2, a whole gamut of decent shells). And other OSs are just as bad - solaris not shipping with bash, nobody shipping with zsh, HPUX having a binary for 'll' but nobody else (aliases only!), the dhcp server on NT doesn't do DHCPINFORM *either* (and it's a hell of a lot harder to port beta 3 from the ISC to NT than it is to any unix)...
So what we have is selective reporting - masking half the facts in order to spread FUD.
As far as standardising on a distribution goes, I don't think it's the best solution. People should be able to run whatever they like, and everyone should know how to operate stuff, or not try to use it. Example: I run Debian on here, two folks run RedHat and we don't have any SuSE users yet. The main unix server boxes are in a mess because HP-ignorant twerps will play with HP boxes, Solaris-ignorant twerpettes with solaris boxes. If anyone dares to touch my linux box, they'd get a severe beating for trying to use 'rpm' instead of apt. BUT the fault is not with me for using the 3rd-most popular distro, but with those who don't know how to use it themselves, but presume "oh it's linux", "oh it's unix".
Other things: there are productivity packages around (I don't believe in them - they're too Dilbert/PHB-ish, but they exist), and there's always StarOffice (been running for 3 days non-stop on here!).
"Scalability of its support model"? Does that by any chance translate as, "I don't understand Usenet ettiquette and would rather pay megabucks per annum for someone else to come round and fix it"? How much contempt am I allowed to exhibit before being marked down as flamebait?;)
There ought to be an HTML verson of "don't feed the trolls!" by now, just for this...:)
(Just to clarify, the parent comment to this is a followup to someone else's, not an original.)
Otherwise: I noted that one of the goals was to 'encourage women to persue engineering'. Initial thought: "oh no, someone thinks there have to be more females in Engineering". Second thought: better that they encourage women to take part in engineering, than that they tell engineers to make their profession more female-friendly. As a general point on this 'equal opportunity' discrimination lark, I'd *much* prefer things to be this way round - rather than arbitrarily imposing ideas like 'must be 50-50 male/female', it should be acknowleged that guys & gals are attracted to different things differently. So, that said, good on them for getting there with the satellite!
Next thought: I'm still searching for a one-liner purpose for the "mission". It seems to have a circular purpose that it goes up there, broadcasting the website (URL? Content?) so that radio amateurs can go visit it. When they get there, check the 'mission' link, they get something that looks to be saying "we put this here so you'd find it"... erm...
Oh yeah. If it's a team of only "girl geeks", what's someone with a name of Duncan being involved in the computer engineering side?;)
German differs between German and German!:8) (Yup, I've seen it happen where folks from Munich and near Hannover have slowed down so much that *I* could understand most of what they were saying, where certainly the more northern stuff at full speed left me well behind). It was quite fun:)
As an aside: I've just acquired the soundtrack from 'The Matrix' which Rammstein's 'Du Hast', including (Bavarian) abbreviation of 'habe' into 'hab''. Babelfish can't cope with this at all, and even confuses 'nichts' into 'anything'...
As far as regionalisations go, I don't know that many Yorkshire natives who use 'thee' and 'thou', at least of the current generation. (And that's where I come from;) I'll also go one further, and suggest that regionalisations *must* remain as intact as possible. (Although I won't be asking version one to translate 'there'll be trubble up at' t'mill';)
Am I the only one so far to get bad vibes of Tower of Babel meets 1984?;)
Not only do they not know what they're doing with regard to other OSs, they've obviously never even beend round SysInternals - as there's a FAT32 driver right there, for NT...
"If you do not know your administrator account password, you will have to completely reinstall Windows NT" - what crap! Who *doesn't* run their personal NT boxes as an administrative user, who can blank the Administrator account's password any time they like? (OK, apart from those of us who run linux instead all day..:) As for using a naff pagefile size - doesn't it grow? Can't users cope with error messages about running low / out of virtual memory? If not, then it's a design fault as much as anything else. But I'd prefer to put it down to typical luser stupidity.
Altogether, more FUD - albeit against microsoft today.
What does this kind of article actually achieve? Those of us "in the know" dismiss these things as crap spouted by someone with keyboard diarrhoea, those who don't know what they're doing won't understand a word of it anyway. Seems pretty pointless to me!
Coo. If short-sightedness is something that's, at least in part, a conditioning / environmental thing, how does the effect of external circumstances change over time? (Ie, at what age is it "safe" to read?:)
ISTR when I first got a PC, deliberately with a 17" monitor, I was pulling 17hr working days doing nothing but look at it. It was actually slightly relaxing on the old optics. (By contrast, I found sitting in my *own* swivel-chair sea-sickening, despite having sat in assorted swivel-chairs long before....) Now I have a 22" screen at work and a 15" thing at home, and I really don't mind running even up to 1280x1024 on the 15" and 1600x1200 on the 22" - the latter all day, although not necessarily all that close up. Aged 24, is this going to be even remotely a problem?:)
It doesn't have to be tarball - FWIW quake2 comes in RPM format, but with a static build of rpm2cpio on the CD and (IIRC) the install script checks to see whether to use rpm or rpm2cpio...
I think there's a fundamental problem with these things, certainly one that restricts my predicted usage of them:
a) of course I don't want M$loth to know one word of data about *me*, let alone have any control over my finances b) if it's implemented client-side, as a cookie, then I use far too many browsers in the average day (Netscape and Konqueror under linux, IE & netscape under Windoze) both at home and at work, for it to be viable c) if it's implemented client-side as a browser extension, then the chances are I'll not want to use any browser that's been so mangled as to have proprietary extensions in it - open-source standards as approved by the W3C, or it's not a web browser.
So - I guess I'll just have to be one of the money-havenots, or something. Bummer!
Has anyone thought that half the problem comes from the phrase "open the mail", not necessarily the mail itself? I treat it as indicative that people want, and are given, flashy features such as (for example) javascript-enabled mail clients (netscape), which then prove to have problems.
If we were to give up on the verb "open", and actually *read* mail instead - insert "hey everyone let's use mutt" rant here - then would we have the same problem? I think not.
Now, how do we persuade people to use simple mail clients that actually do just what they need with *NO* fancy features? "Look! PigMail has new security plugin! Complete with rm/var/spool/mail/$USER technology!"
Perhaps we ought to bear in mind that it's only expected to be amassing 40Gb a year - it's hardly as though you're going to want to do a select * across one table that size and get all the events for a year. So yes, journalling would be a good idea, but it would also be very sensible to invest in a regular (weekly? daily?) archival solution. He could always use postgresql, pg_dump | gzip, and then drop the tables, every night:)
Yeah. The other thing that I remember from the page is thinking, the ODBC 3 results are somewhat screwy - it depends strongly on what ODBC driver you go through, as the rule with ODBC is not 'if the database supports it, we do' but rather 'if it's in the spec and we choose to implement that feature, you can then look up its availability and use it if you choose'. Rather long-winded but it sums up feature support in ODBC fairly well, I think. What I'm saying is, there are one or two bits & pieces in the results that I believe are artifacts of the chosen ODBC driver, not of postgesql (which is the one I was looking at most).
And yes, I'd expect some confusion over ODBC's DATETIME type being merely 'DATE' to Oracle, especially if the script was written to look for only one-to-many relationships the wrong way round, as it were.
That is an alternative, although if there's one thing you might want in a high-load network-management station, it's transaction support.
From the PostgreSQL Admins' guide (http://www.postgresql.org/docs/admin/manage-ag175 6.htm#AEN1774), I note there's a section on large tables, including the line: "Since Postgres allows tables larger than the maximum file size on your system, it can be problematic to dump the table to a file, since the resulting file will likely be larger than the maximum size allowed by your system. As pg_dump writes to stdout, you can just use standard *nix tools to work around this possible problem: Use compressed dumps: % pg_dump dbname | gzip > filename.dump.gz" etc.
As regards postgresql being "as grown up as Oracle", erm, it's certainly a viable solution, and doesn't suffer from the bloatware problems of Oracle. However, you might well find that it's not as fast as Oracle or MySQL for adding lots of events as rows in the table quickly - it all depends. Try it out and see:)
An alternative option: 5) they're trying to make a commercial distribution, that corporate mindset will worm its evil^Wnon-GPL way from Intel to the top brass at TurboLinux.
You might gather I'm not entirely in favour of commercial alliances, only friendly ones...
Anyway. I tried Turbolinux a while ago, and it was the only thing I've ever had diald and Freeserve working together on (nowadays I don't bother - Debian, pon and poff work well enough!). It wasn't bad, it was version 2.0 ish, and it was RedHat with various deviations like all the turboXXXconfig applets. Oh, and ISTR it set me up with afterstep by default. Nothing spectacular, nothing too awful...
I guess we'll just have to see what comes of the Intel thing...
From ftp://ftp.gnu.org/pub/gnu/GPL , I think you're pretty much exactly right. The requirement is, as others have already posted, that the source is either distributed with it, or made available and you know that's the state of play, etc.
Of course, in practice, it would be rather nice if kernel mods made their way back to Linus or somebody else in the kernel team for approval...
I've not hacked java particularly, but from such as I know...: The object file doesn't have 'lines'... it's tokenised, binary.
It's a two-stage thing: you write in java, which obviously looks similar to C++ source, to the not-well-trained eye. At least it's plain text at this stage. Then you compile it into some messy looking.class format. This contains all the same symbol names as the real source, but the whole file is complete garbage to even attempt to understand. The machine (JVM) itself reads this binary stuff and interprets it - binary encoding of token by binary encoding of token. So there's a fairly simple mapping between the instructions you gave and the things the interpreter phase of it does.
So.java is source to us,.class we regard as object. From the JVM's PoV,.class is source, actions are the results.
This isn't all that different to ordinary java.class files, is it? I mean, that's an intermediary tokenised format, you don't have to use a JIT compiler on it... and you can mangle the symbol names (maybe not removing the idea of symbol names entirely, but as good as)...
If there's one thing it *would* achieve, it'd possibly help introducing the government to the concept of "brain" - they really need to stop talking beaurocratic crap and to produce laws that actually talk about the technology in the correct terminology, for starters. Half the problem at the moment is that legalese is not slashdot-speak, I think.
Slashdot Mirror
Why do people criticise, say, checkpoint firewalls, for giving out the fact that they are checkpoint firewalls, when you telnet into them?
The point is not just "ooh look, there's a port open" but rather that the information has been let loose, "so goodness knows what's going to happen". (There's a reason I didn't let on anything particularly interesting with my two URL links!)
So there's a firewall in the way (?solaris?iss?) and there are non-80 ports that are filtered, while 80 is the only one open...
Sure, maybe the worst that can happen is a remote DoS attack - quite possibly the case, but I'm not going to launch *that* - but do you really want the Royal webpages getting defaced and/or being unavailable, even for half an hour?
RedHat 5.2 is ancient, apache 1.3.3 is buggy (insecure? quite possibly)... so thanks for the information!
Just an idea... I wouldn't have let loose what it was running on, myself - but then again, it would be interesting to see if they have crack-attempt logs in the manner of antionline, and so on...
Why?
Not *everyone* sweats away listening to dodgy loud metal trying to get the latest patch to apache to work, or whatever they're hacking on, for free. Businesses still exist and like to make money - it's way too early for us all to go broke & hippy about it all.
Admittedly, M$loth apps are the last things I want to run - I do, however, need NT at home for the scanner and quake2, but otherwise I work off linux all day - because my ideology is not fundamentalist "linux has to be better", but rather, it *is* the way I want to work.
I generally agree.
The problem is "1 window manager, 1 desktop environment, 1 version" - we all know that with that limitation, as some would have it be, OpenSource goes out t'window.
Open-Source is not linux, in some ways it's "more than" linux - it's the software development ethic, more than anything else. And FreeBSD (and others) are, as yet, unpolluted by the influx of marketroids.
Is it possible that one day, we'll all be saying "Open-Source is essential" and actually wake up & realise it doesn't happen any more, given the rate at which non-free things are being ported?
I've not done IE under unix (of any sort) recently, so can't update you on that.
But a comment on the browser scene: netscrape, sure, is underfeatured and over-bloated. IE is evil, of course. Mozilla isn't remotely ready for "public consumption".
My recommendations on the unix browser scene boil doing to either Opera (as & when they get their act together) or Konqueror - part of the KDE. I'm using the latter now as one of my mainstays for non-javascript browsing - even slashdot comes out OK in it, and it's stable enough (1.1.2). Oh, and it also renders PNGs correctly in a webpage.
Yes indeed - so how come we have lots of different ways of slagging them off all marked up as "insightful" and various other things, and yet we're no good at praise where it's due?
Sure it was the wrong thing to start with, but at least Corel listened to the community, or something, and have now rectified it. This is not bad - it's not necessarily as good as it could be, but it *does* show "something stirs", or whatever.
It's a misfeature of slashdot, IMO, that there is neither room for duplication of feeling nor for going that far off-topic.
Well, I don't know why everyone else on here is into heavy / metal / noisy type stuff (says me, choosing terminology so's not to offend :) but as for me, I like a nice bit of quiet RunRig or Wolfstone, for coding, newsgroups, ...whatever.
:)
Must be mad
Yeah, in that case, the desktop environment differences do cause some problems - at least 3 different window managers (enlightenment, kwm and wm) are popular with their assorted CORBA-using layers above (KDE, Gnome). ;)
/usr/local/so51/ or similar is suggested, with ensuing wars about /usr/local and /usr and /opt and all that stuff. :)
However, it is quite possible to run several of these all at once - I've got KDE (with kwm) and a gnome panel sitting here quite happily. The fun & games only really start with gmc running as it's anyone's guess who gets the desktop clicks
As far as the average user approaching unix, linux, or X goes, there's bound to be some confusion. The term 'window manager' must be explained, at the very least. What we can't have is the sort of state of play, "what window manager are you using?" - "linux!" *duh*...
And while I'm not exactly the best support / trainer around, if the 30-yr-old unix model with X and stuff can finally take off, all to the good.
Oh, as for star office, I have actually seen folks asking where it should be installed - normally
So there are users out there who want to do The Right Thing(TM), fortunately. And these I have a chance of getting on favourably with...
"True: Linux may not have many standards, but you can standardise on a distrobution. Decide on what one you think is best for your operation then stick with it."
;)
:)
The problem is, that's not true. We have RFCs, just like everything else. Some stuff implemented on linux doesn't live up to the RFCs in question (dhcpd 2, for example, doesn't implement DHCPINFORM), but then again it also implements some stuff far better than mere standards-compliance (gnutar, bzip2, a whole gamut of decent shells).
And other OSs are just as bad - solaris not shipping with bash, nobody shipping with zsh, HPUX having a binary for 'll' but nobody else (aliases only!), the dhcp server on NT doesn't do DHCPINFORM *either* (and it's a hell of a lot harder to port beta 3 from the ISC to NT than it is to any unix)...
So what we have is selective reporting - masking half the facts in order to spread FUD.
As far as standardising on a distribution goes, I don't think it's the best solution. People should be able to run whatever they like, and everyone should know how to operate stuff, or not try to use it.
Example: I run Debian on here, two folks run RedHat and we don't have any SuSE users yet. The main unix server boxes are in a mess because HP-ignorant twerps will play with HP boxes, Solaris-ignorant twerpettes with solaris boxes. If anyone dares to touch my linux box, they'd get a severe beating for trying to use 'rpm' instead of apt. BUT the fault is not with me for using the 3rd-most popular distro, but with those who don't know how to use it themselves, but presume "oh it's linux", "oh it's unix".
Other things: there are productivity packages around (I don't believe in them - they're too Dilbert/PHB-ish, but they exist), and there's always StarOffice (been running for 3 days non-stop on here!).
"Scalability of its support model"? Does that by any chance translate as, "I don't understand Usenet ettiquette and would rather pay megabucks per annum for someone else to come round and fix it"? How much contempt am I allowed to exhibit before being marked down as flamebait?
There ought to be an HTML verson of "don't feed the trolls!" by now, just for this...
(Just to clarify, the parent comment to this is a followup to someone else's, not an original.)
;)
Otherwise: I noted that one of the goals was to 'encourage women to persue engineering'. Initial thought: "oh no, someone thinks there have to be more females in Engineering". Second thought: better that they encourage women to take part in engineering, than that they tell engineers to make their profession more female-friendly.
As a general point on this 'equal opportunity' discrimination lark, I'd *much* prefer things to be this way round - rather than arbitrarily imposing ideas like 'must be 50-50 male/female', it should be acknowleged that guys & gals are attracted to different things differently.
So, that said, good on them for getting there with the satellite!
Next thought: I'm still searching for a one-liner purpose for the "mission". It seems to have a circular purpose that it goes up there, broadcasting the website (URL? Content?) so that radio amateurs can go visit it. When they get there, check the 'mission' link, they get something that looks to be saying "we put this here so you'd find it"... erm...
Oh yeah. If it's a team of only "girl geeks", what's someone with a name of Duncan being involved in the computer engineering side?
Since when was 'woman' a different sex to 'girl'?
:P
And don't be offensive to us pigs, either
As an aside, your quote has a non-optimal comma. If I like computers, invade other people's privacy and take away their networks? ;)
And of course I wish they'd use the correct terms, 'cracker' and 'nerd'...
German differs between German and German! :8) :)
;) ;)
;)
(Yup, I've seen it happen where folks from Munich and near Hannover have slowed down so much that *I* could understand most of what they were saying, where certainly the more northern stuff at full speed left me well behind). It was quite fun
As an aside: I've just acquired the soundtrack from 'The Matrix' which Rammstein's 'Du Hast', including (Bavarian) abbreviation of 'habe' into 'hab''. Babelfish can't cope with this at all, and even confuses 'nichts' into 'anything'...
As far as regionalisations go, I don't know that many Yorkshire natives who use 'thee' and 'thou', at least of the current generation. (And that's where I come from
I'll also go one further, and suggest that regionalisations *must* remain as intact as possible. (Although I won't be asking version one to translate 'there'll be trubble up at' t'mill'
Am I the only one so far to get bad vibes of Tower of Babel meets 1984?
Not only do they not know what they're doing with regard to other OSs, they've obviously never even beend round SysInternals - as there's a FAT32 driver right there, for NT...
:)
"If you do not know your administrator account password, you will have to completely reinstall Windows NT" - what crap! Who *doesn't* run their personal NT boxes as an administrative user, who can blank the Administrator account's password any time they like? (OK, apart from those of us who run linux instead all day..
As for using a naff pagefile size - doesn't it grow? Can't users cope with error messages about running low / out of virtual memory? If not, then it's a design fault as much as anything else. But I'd prefer to put it down to typical luser stupidity.
Altogether, more FUD - albeit against microsoft today.
What does this kind of article actually achieve?
Those of us "in the know" dismiss these things as crap spouted by someone with keyboard diarrhoea, those who don't know what they're doing won't understand a word of it anyway. Seems pretty pointless to me!
Coo. :)
:)
If short-sightedness is something that's, at least in part, a conditioning / environmental thing, how does the effect of external circumstances change over time? (Ie, at what age is it "safe" to read?
ISTR when I first got a PC, deliberately with a 17" monitor, I was pulling 17hr working days doing nothing but look at it. It was actually slightly relaxing on the old optics. (By contrast, I found sitting in my *own* swivel-chair sea-sickening, despite having sat in assorted swivel-chairs long before....)
Now I have a 22" screen at work and a 15" thing at home, and I really don't mind running even up to 1280x1024 on the 15" and 1600x1200 on the 22" - the latter all day, although not necessarily all that close up.
Aged 24, is this going to be even remotely a problem?
It doesn't have to be tarball - FWIW quake2 comes in RPM format, but with a static build of rpm2cpio on the CD and (IIRC) the install script checks to see whether to use rpm or rpm2cpio...
I think there's a fundamental problem with these things, certainly one that restricts my predicted usage of them:
a) of course I don't want M$loth to know one word of data about *me*, let alone have any control over my finances
b) if it's implemented client-side, as a cookie, then I use far too many browsers in the average day (Netscape and Konqueror under linux, IE & netscape under Windoze) both at home and at work, for it to be viable
c) if it's implemented client-side as a browser extension, then the chances are I'll not want to use any browser that's been so mangled as to have proprietary extensions in it - open-source standards as approved by the W3C, or it's not a web browser.
So - I guess I'll just have to be one of the money-havenots, or something. Bummer!
Has anyone thought that half the problem comes from the phrase "open the mail", not necessarily the mail itself?
/var/spool/mail/$USER technology!"
I treat it as indicative that people want, and are given, flashy features such as (for example) javascript-enabled mail clients (netscape), which then prove to have problems.
If we were to give up on the verb "open", and actually *read* mail instead - insert "hey everyone let's use mutt" rant here - then would we have the same problem? I think not.
Now, how do we persuade people to use simple mail clients that actually do just what they need with *NO* fancy features?
"Look! PigMail has new security plugin! Complete with rm
;]
Perhaps we ought to bear in mind that it's only expected to be amassing 40Gb a year - it's hardly as though you're going to want to do a select * across one table that size and get all the events for a year. :)
So yes, journalling would be a good idea, but it would also be very sensible to invest in a regular (weekly? daily?) archival solution.
He could always use postgresql, pg_dump | gzip, and then drop the tables, every night
Yeah. The other thing that I remember from the page is thinking, the ODBC 3 results are somewhat screwy - it depends strongly on what ODBC driver you go through, as the rule with ODBC is not 'if the database supports it, we do' but rather 'if it's in the spec and we choose to implement that feature, you can then look up its availability and use it if you choose'.
Rather long-winded but it sums up feature support in ODBC fairly well, I think.
What I'm saying is, there are one or two bits & pieces in the results that I believe are artifacts of the chosen ODBC driver, not of postgesql (which is the one I was looking at most).
And yes, I'd expect some confusion over ODBC's DATETIME type being merely 'DATE' to Oracle, especially if the script was written to look for only one-to-many relationships the wrong way round, as it were.
That is an alternative, although if there's one thing you might want in a high-load network-management station, it's transaction support.
5 6.htm#AEN1774), I note there's a section on large tables, including the line:
:)
From the PostgreSQL Admins' guide (http://www.postgresql.org/docs/admin/manage-ag17
"Since Postgres allows tables larger than the maximum file size on your system, it can be problematic to dump the table to
a file, since the resulting file will likely be larger than the maximum size allowed by your system.
As pg_dump writes to stdout, you can just use standard *nix tools to work around this possible problem:
Use compressed dumps:
% pg_dump dbname | gzip > filename.dump.gz"
etc.
As regards postgresql being "as grown up as Oracle", erm, it's certainly a viable solution, and doesn't suffer from the bloatware problems of Oracle.
However, you might well find that it's not as fast as Oracle or MySQL for adding lots of events as rows in the table quickly - it all depends.
Try it out and see
An alternative option:
5) they're trying to make a commercial distribution, that corporate mindset will worm its evil^Wnon-GPL way from Intel to the top brass at TurboLinux.
You might gather I'm not entirely in favour of commercial alliances, only friendly ones...
Anyway. I tried Turbolinux a while ago, and it was the only thing I've ever had diald and Freeserve working together on (nowadays I don't bother - Debian, pon and poff work well enough!). It wasn't bad, it was version 2.0 ish, and it was RedHat with various deviations like all the turboXXXconfig applets. Oh, and ISTR it set me up with afterstep by default.
Nothing spectacular, nothing too awful...
I guess we'll just have to see what comes of the Intel thing...
From ftp://ftp.gnu.org/pub/gnu/GPL , I think you're pretty much exactly right.
The requirement is, as others have already posted, that the source is either distributed with it, or made available and you know that's the state of play, etc.
Of course, in practice, it would be rather nice if kernel mods made their way back to Linus or somebody else in the kernel team for approval...
I've not hacked java particularly, but from such as I know...:
.class format. This contains all the same symbol names as the real source, but the whole file is complete garbage to even attempt to understand.
.java is source to us, .class we regard as object. From the JVM's PoV, .class is source, actions are the results.
:)
The object file doesn't have 'lines'... it's tokenised, binary.
It's a two-stage thing: you write in java, which obviously looks similar to C++ source, to the not-well-trained eye. At least it's plain text at this stage.
Then you compile it into some messy looking
The machine (JVM) itself reads this binary stuff and interprets it - binary encoding of token by binary encoding of token. So there's a fairly simple mapping between the instructions you gave and the things the interpreter phase of it does.
So
Is that good enough?
This isn't all that different to ordinary java .class files, is it?
I mean, that's an intermediary tokenised format, you don't have to use a JIT compiler on it... and you can mangle the symbol names (maybe not removing the idea of symbol names entirely, but as good as)...
If there's one thing it *would* achieve, it'd possibly help introducing the government to the concept of "brain" - they really need to stop talking beaurocratic crap and to produce laws that actually talk about the technology in the correct terminology, for starters. Half the problem at the moment is that legalese is not slashdot-speak, I think.