Re: [SLUG] Vista .. anti-Linux ?
On Sat Apr 29, 2006 at 14:20:28 +1000, [EMAIL PROTECTED] wrote: Benno: On Fri Apr 28, 2006 at 20:18:15 +1000, Malcolm V wrote: On Friday 28 April 2006 19:55, Adam Bogacki wrote: snipped http://www.theregister.co.uk/2006/04/27/schneier_infosec/ Getting back to the topic, I believe that it is possible for a system to detect whether it has been chain-loaded from some other bootloader and then refuse to run if it detects this. The system only works off the officially sanctioned bootloader and this bootloader never boots anything else -- no more dual boot. Probably makes it harder to use MS libraries in wine, also might kill Xen, VMware and all those handy tools that give you a chance to make a few MS-Windows licenses go a long way... Does this give any better security than a well-known encryption algorithm (e.g. AES) plus a passphrase plus a key device (e.g. USB, etc)? No it doesn't, it is probably worse because if your motherboard chip dies you won't be able to recover your data on a different motherboard. That means you have to have an unencrypted backup which in turn becomes the weak point. And from the BitLocker tech article on the MS website, it appears to have a way of working in exactly the mode you describe. Plus its optional. So, its only going to be a problem, if you choose to use Vista, and then choose to enable Vista, and then choose to work in the TPM mode. (And I'm not convinced you couldn't setup the TPM such that you say you trust a particular chain loader configuration, and I'm sure if it is possible, and people want this, then someone will make it easy to do.) Of course this could be seen as scary from a what could they do next, point of view. E.g: to view some media you need to be running Vista and need remote attestation that requires you to use have TPM enabled and then the remote party will only trust a Vista install. Now *that* would be evil. But I think BitLocker itself is a way from that. And of course we could implement the same stuff on Linux, to make it harder for people to use Vista with it. Muhahaha! ;) Benno -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Erik: It should also be easy to prove by now that either Rob or Peter has written more code than you, or anyone you can name that swears off test driven developement. We say that (as long as you ignore genuine knockoffs with only one user) based on our experience, with all other things being equal, test driven development costs less. I find argument by authority a little difficult to reconcile with my notion that the truth or falsity of an idea is founded not so much in where the idea comes from but whether concrete physical examples stand in support of the idea. I am fully aware that there are a number of practical efficiencies in using argument by authority since it saves time and effort spent researching evidence. Then again, methods for taking short-cuts to save time and effort are what this discussion is all about, it seems that you have neglected to provide a test suite with your argument (i.e. verifiable evidence) and have instead chosen to knock something up quickly with little regard for a quality case (i.e. plug the resumes of a few people and propose that if such people support your idea then it cannot be wrong). You no doubt chose to take such a short-cut in the belief that this thread is not of high consequence and thus does not warrant doing a proper job. That's quite a reasonable value judgment, the sort of judgement that people make all the time... a judgement that people apply to technology in all forms and software in particular. As for how much code I've written, sadly I've lost count of the details and it isn't in my blood to enter into a pissing contest either. Even sadder, probably the bulk of what I have written has been gobbled up by history due to a number of factors: [1] Written for a proprietary platform operating system (now obsolete). [2] Written under commercial contract that prevented release of the code (at least 70% of it went here). [3] Solves problems that only a small number of people are immediately interested in. An possibly relevant example from catg [2] above ... See this boat: http://www.solarsailor.com.au/ I wrote the code that spins the motors, it runs on a single chip microcomputer (at least, I wrote the draft that was the first draft to actually make a motor spin, there had been a few drafts before that). Is my name written on the boat anywhere? I probably don't have to answer that. By the way, I also built the first prototypes of the drive and monitoring circuits to make the full controller boards (two copies, one for left drive, one for right), although someone else (whose name is also not written on the boat anywhere) did the theoretical design for that circuit. I don't even have a copy of the code that I wrote, nor the circuit design and if I did I couldn't give it to anyone. That's the reality of commercial engineering for a huge number of people, yes Open Source is starting to change that but it won't change overnight and it won't bring back the past for people who have already lost their own work. If you did manage to get a copy of the code, you couldn't run it unless you had the right chip and even then it wouldn't be much use unless you had a suitable motor drive circuit and a motor too. Writing a test harness for embedded chip code (written in assembler) requires a decent emulator for that chip. There was such an emulator for MS-Windows at the time but it was not in a convenient form to build into a test harness. I did build a miniature motor for testing both the software and the hardware but this was not a regression test nor could it easily be automatically repeated, plus the physical characteristics of the test motor were a very rough approximation of the real motor. A proper test harness with a software simulated motor plus a chip emulator that can be built into a suitable framework would have been a huge cost (no doubt an excellent project but also a slow and expensive one). This is one of many examples of software projects where a regression test suite could be made, would be useful, but is not cost effective for the project. Let's not forget that the entire Solar Sailor project was itself a proof of concept. To my knowledge, only one such boat has ever been built, but for every new idea there must be a first one. Maybe it was a bad idea that never should have been built, maybe the price of oil just isn't high enough yet but that's how things go... You might also look at some of these (catg [3] from above): http://bespoke.homelinux.net/gnucap/r30.pdf http://bespoke.homelinux.com/TCPLAG/ http://cran.r-project.org/src/contrib/Descriptions/gafit.html I'm not terribly bothered by doing things that fall into category [3] because although each one of them is not a huge contribution to the bright lights of mainstream development, they are all projects that interested me and I didn't get paid for any of them so there's no motivation on my part to make them more than what
[SLUG] Serious Printer Question -- Samsung CLP-550N
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Here's a link to Samsung's page: http://www.samsung.com/au/products/printerfaxcopysolutions/printerfaxcopysolutions/clp_550n.asp I'm thinking about buying one of these printers, they are available under $800 and they support Postscript and PCL5c as well as the undocumented (i.e. useless) Samsung Printer Language. They also have a rather low cost-per-page when you add up the consumable costs (note that the KYOCERA FS-C5020N has a lower cost per page but you have to print about 4 pages to break even, when you consider the different up-front costs, I doubt I'll print that many). Note that the above break-even calculation is based on the presumption that both printers come fully loaded with toner when you first buy them... but it's starting to become standard practice to sell printers close to empty. What makes me nervous is that the Samsung is a lower price for equivalent specs to just about every other colour-laser on the market at the moment... is it just good value or am I missing a trick? I'm wondering if they sell it with no colour toner at all and make you buy your first batch up-front (which would bring the price in line with the rest of the market). Has anyone tried one of these? Any inside knowledge in the colour-laser market? I note that LEXMARK C510 looks like a close contender with lower up-front cost and higher consumable cost (but still Postscript and PCL support, woot). Unfortunately I just don't trust the Lexmark name after experience with their disposable inkjet printers. Yes I know that it's not fair to compare a laser with an inkjet but if Lexmark want to put their name on junk then I'll treat them like junk, that's what trade-marks are intended for. - Tel -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.2.2 (GNU/Linux) iQIVAwUBRFMp7sfOVl0KFTApAQKULw/7B6ciMtTtYTVdZ3G7VFqClspUJ4HdMda6 NmhaoHF/ktdyiJ+Y6AWcUQs+9MQt7kZn6bE3az+gsViQ95Yc+t70Kiqct3rJVq/W K/BLRbZiZ3IxCcAJTX1Fi/tsGDUNTMG/u/Zkzq9Q80Cze3IZGGtjpmQvaJNELwu8 qCj/ER+Iq72rBqymURBvgApEV+19dTnSDSFhqSK9JpPvHkh37MsCjnTR+uaUS/xT HK+8CvXCErcvSsLUmotTLZpVLt2e2f7PjFLl61fypl/m6wQeYsHjzbZqTpZtwBjh vRUTbzwGFSDF2nABDb+7VX9lPrXMp9fto4Emu1VXB1fzpk6Cd+uXrIvWj9egSmeC nJ0A5gUr9Jx6tKPQstDWUCR2i+owKxvJA5srE37klq7ufdqaPHiMofJL1jkV2e2w KVzGYC5Du1QugQFDlF86ELRTT3wcSpSzdVDbaqZgF92MorfsQtL3xeb7jWcsFH9w ZVZTuvNViH05Pqfocd0RijdPnS4eMZ1s4KD9KBwJw6yqkvp/enrkxtss+r2TKrUk WIdgvWzk0hhhlpEJxpUwCDbo5iEbXS1MKTco05oZC8o3BqxisKJznWus1M0Qp1Qi S45FFr4/kCRufZg1EKv8dojpoTI/KI4To9kHQFToivja8R+aryTNSsOqls7N9WS9 FnBhEUhmnsc= =Yujh -END PGP SIGNATURE- -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
On Saturday 29 April 2006 16:11, [EMAIL PROTECTED] wrote: With regards to last night's Slug meeting and using automated testing, I think everyone agrees that writing (and using) test cases produces higher quality code with less bugs. My point is that higher quality output doesn't come for free, it requires effort and that usually means someone has to pay for it. If you are writing code because you want to write code and you want to produce the best result that you can then sure you are going to want to put the extra time into it and do work you can be proud of. Use whatever techinique you think will work, regression tests will help, good design documentation will help, feedback from users helps too. On the other hand, if you want to get a job done, get paid and get out of there (and that's the way 90% of business works, sorry to say) then the fact is that corners get cut and the end result is not high quality. With Open Source style programming, it might eventually end up as high quality software once enough people have got interested enough to bash it into shape (and often after the second or third re-write from scratch) but it very rarely starts out that way. In the early days of a project it is hard enough to get enough time and effort in to make it work at all (even as a buggy proof of concept) let alone produce a high quality masterpiece. [snip] I've stood on both sides of the fence. Doing a Software Specification and Software Design Document is often a formal utter waste of time (OK it makes the QC people happy). The reason that software gets written 3 times is that it produces the BEST and HIGHEST QUALITY code. It just takes the longest time to do! With a working copy you can see the shortcomings, see the better implementation and (It frustrated the living wosname out of my software manager) write the SSD and SDD (software spec and design docs) after the software is written. My favourite eg: PTC track (rail) transponder readers. I watched the 'expert consultant' spend 3500 hours, watched his tries and fumbles and finally fail the official 'type testing' acceptance. In 30 days I wrote the transponder reader, passed the type testing, and 15 years later on 100s trains every day throghout NSW over 100s transponders there are (and have been) no defects! Gave the source to my manager and said write the SSD and SDD now! So IMHO the formal QC documentation paradigsm produces 'acceptable' code in the shortest time, takes MUCH longer to produce excellence. Write/rewrite 3 times is a quicker way to excellence. You want it fast, good and cheap ... James -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Serious Printer Question -- Samsung CLP-550N
On Saturday 29 April 2006 18:55, [EMAIL PROTECTED] wrote: Here's a link to Samsung's page: http://www.samsung.com/au/products/printerfaxcopysolutions/printerfaxcopyso lutions/clp_550n.asp I'm thinking about buying one of these printers, they are available under $800 and they support Postscript and PCL5c as well as the undocumented (i.e. useless) Samsung Printer Language. They also have a rather low cost-per-page when you add up the consumable costs (note that the KYOCERA FS-C5020N has a lower cost per page but you have to print about 4 pages to break even, when you consider the different up-front costs, I doubt I'll print that many). Note that the above break-even calculation is based on the presumption that both printers come fully loaded with toner when you first buy them... but it's starting to become standard practice to sell printers close to empty. What makes me nervous is that the Samsung is a lower price for equivalent specs to just about every other colour-laser on the market at the moment... is it just good value or am I missing a trick? I'm wondering if they sell it with no colour toner at all and make you buy your first batch up-front (which would bring the price in line with the rest of the market). Has anyone tried one of these? I have one, I believe it to be one of the best value colour laser printers available. It comes with a full set of toner cartridges which appear to be full. Well I just had to replace the cyan toner which had done about 5,000 pages I think. The other cartridges have yet to be replaced. Having a duplexer built in makes dual page printing a dream. To install the software is one of the easiest I've come across which intergrates with CUPS. -- Regards, Graham Smith -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
On Sat, 2006-04-29 at 16:55 +0800, [EMAIL PROTECTED] wrote: My favourite eg: PTC track (rail) transponder readers. I watched the 'expert consultant' spend 3500 hours, watched his tries and fumbles and finally fail the official 'type testing' acceptance. In 30 days I wrote the transponder reader, passed the type testing, and 15 years later on 100s trains every day throghout NSW over 100s transponders there are (and have been) no defects! Gave the source to my manager and said write the SSD and SDD now! So IMHO the formal QC documentation paradigsm produces 'acceptable' code in the shortest time, takes MUCH longer to produce excellence. Write/rewrite 3 times is a quicker way to excellence. What I find interesting here is that *neither* of two development methodologies you describe here are test-first development... and thus not comparable to what my talk [hopefully] encourages people to do. wikipedia has a nice little article on this too : http://en.wikipedia.org/wiki/Test_driven_development. In the context of my talk, your expert consultant was clearly not using a test suite to prevent regressions, and I classify what you did as 'ad hoc human testing'. Because you appear to not have written any automated tests, if there is a design change today, they will be unable to tell in a cheap, automated, fashion if the required code changes break any functionality or not. I have not, claimed that one *cannot produce defect free software* without doing test-first or test-driven-design in the process. Thats clearly a fallacy. However, I do claim, and there are many anecdotal references 'out there' on the web that back this claim up, that a programmer following a test-first discipline is more productive than the same programmer not following the same discipline. As for the time scale that this productivity can be measured on, *once the skill is learnt*, the productivity difference is visible on timescales as short as a day. Someone learning the discipline - still learning or integrating the splinter skills required to successfully practice this - will be less productive than they are when they do what they are already used to. But thats the same curve as learning any new skill. Once the skill has been learnt (and its a *bitch* to learn without a mentor: its at least 10-20 times longer to pickup from books and examples than from pair programming with a practitioner). I covered the things that become easier, and why during my talk. I plan to follow up the talk with material (perhaps a talk, perhaps paper[s]) aimed at making the TDD knack easier to learn without a mentor. Unfortunately, Telfords scepticism is entirely reasonable: stating that TDD is easier doesn't convey the knowledge that it is. And its like other composite skills : until the synthesis is experienced, its almost impossible to communicate the depth of change it brings. Telford, as I understand your argument, it boils down to: The folk that pay the bills will not pay for me to write tests because the act of writing those tests will delay the deliverable of the required functionality and as such raise the price. Is that a fair statement? Assuming that it is I happily acknowledge that bill-payers will not accept paying more for the product because the development team decide to do TDD. I'm not debating the validity of that part of your argument. The part of your point I'm debating is the assumption in that 'using TDD increases the total number of programmer hours to deliver a given function-point to the customer.' I think the crux of the disagreement lies in my assertion that the time cost of writing tests first is offset by the increased developer efficiency within the same programming cycle. I've asserted above that a single day is long enough to demonstrate this comprehensively. An I believe that anyone who has invested the time to grok TDD will agree with me on that, and anyone who hasn't will be skeptical :). That said, if one can write *any* automated tests for the code being written, even if you cant test everything (there is always a pragmatic aspect in testing, how much, where, when etc) those things one does write tests for bring the previously discussed benefits. In short, development methodology changes the resourcing equation: its -not- a zero sum resource allocation problem between 'fast, good and cheap'. Rob -- GPG key available at: http://www.robertcollins.net/keys.txt. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Ruby Sunday?
Tomorrow is another Sunday. It does not have to be 'yet' another Sunday when you can get with Ruby! Darling Harbour is action stations at the Photo Imaging Expo. Two done, one to go, see excellent photos (just judged on Friday). And a two halls full of photo-stuff. $20 at the door. (Was free via internet rego before Friday). After, drop in on the Ruby session at Govirtual/Cohi Bar. Up the escalator at the Southern end of Harbourside. High speed wireless 'on tap'. Around 3pm. Don't forget to visit the Cebit site for free rego. http://www.cebit.com.au/ --- Govirtual.com.au/KEYS/[EMAIL PROTECTED] Govirtual is protected by strong antiVirus Spam blocking. You have been whitelisted. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
On Sat, 2006-04-29 at 15:08 +1000, Erik de Castro Lopo wrote: Don't you think we may be right? Absolutely. I have written perl modules with test scripts and without. The ones with test scripts have always been faster to develop. This sounds absolutely wrong however it is what I have found. Yes I work as a programmer under pressure. So this is not a personal exercise. Thanks Ken Foskey FOSS developer -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
On Sat, 2006-04-29 at 19:34 +1000, Robert Collins wrote: wikipedia has a nice little article on this too : http://en.wikipedia.org/wiki/Test_driven_development. There is a step missing from this document. We have to remember to test the test code a little bit. Write test. *** Run test, ensure that it fails. *** write code to make test work. run test, repeat until code works. -- Ken Foskey FOSS developer -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
On Sat, 2006-04-29 at 21:40 +1000, Ken Foskey wrote: On Sat, 2006-04-29 at 19:34 +1000, Robert Collins wrote: wikipedia has a nice little article on this too : http://en.wikipedia.org/wiki/Test_driven_development. There is a step missing from this document. We have to remember to test the test code a little bit. Write test. *** Run test, ensure that it fails. *** write code to make test work. run test, repeat until code works. Another variation of test-driven development requires the programmer to first fail the test cases. The idea is to ensure that the testcase really works and can catch an error. Once this is shown, the normal cycle will commence. This is one of the more popular variations and has been coined the Test-Driven Development Mantra, known as red/green/refactor where red means fail and green is pass. From that document :). Rob -- GPG key available at: http://www.robertcollins.net/keys.txt. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Question on ownerships and permissions of subversion respositories
Hi all I have a question on ownerships and permissions of subversion respositories. I'm setting up a subversion repository to be used by a few users. I therefore didn't want to put the repository in some /home/somebody/ but into /var/lib/ i.e. next to where cvs respositories reside. This is a Debian system. This is what I have done so far: As root create the respository: # svnadmin create /var/lib/svnrepos The users are members of the src group so change the root:root ownership to root:src /var/lib# chown -R root:src svnrepos/ The project is in a TEMP directory so import it from there: TEMP$ svn import . file:///var/lib/svnrepos/project1 -m initial import svn: Unable to open an ra_local session to URL svn: Unable to open repository 'file:///var/lib/svnrepos/project1' svn: Berkeley DB error while opening environment for filesystem /var/lib/svnrepos/db: Permission denied Of course we have to change the permissions of /var/lib/svnrepos This is what they were originally: $ ls -l /var/lib/svnrepos/ -rw-r--r-- 1 root src 379 Apr 29 21:36 README.txt drwxr-xr-x 2 root src 4.0K Apr 29 21:36 conf drwxr-xr-x 2 root src 4.0K Apr 29 21:36 dav drwxr-sr-x 2 root src 4.0K Apr 29 22:05 db -r--r--r-- 1 root src2 Apr 29 21:36 format drwxr-xr-x 2 root src 4.0K Apr 29 21:36 hooks drwxr-xr-x 2 root src 4.0K Apr 29 21:36 locks Looking at the above I decided that as I'm not using webdav and format and hooks/ and conf are probably not modified by users that I don't need to change the ownerships of everything and that indeed might be bad security wise so I just changed db/ and locks/ so that the src group could create or change files in there: From within /var/lib/svnrepos/ # chmod -R g+w db/ # chmod -R g+w locks/ Now the import works fine: $ svn import . file:///var/lib/svnrepos/project1 -m initial import QUESTION 1. Is this the correct thing to do if one wants repositores in /var/lib? 2. Am I lowering security by the chowns and chmods? 3. What's general practice on where developers put their respositories in Debian systems? At present this is on my laptop but it will be done on an internet connected virtual server soon, I'm just documenting everything that I need to do to get it all working correctly and securely. I'll be using svn+ssh. Mike -- Mike Lake Caver, Linux enthusiast and interested in anything technical. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
[EMAIL PROTECTED] wrote: Erik: It should also be easy to prove by now that either Rob or Peter has written more code than you, or anyone you can name that swears off test driven developement. We say that (as long as you ignore genuine knockoffs with only one user) based on our experience, with all other things being equal, test driven development costs less. I find argument by authority You quote that and call it argument by authority? I think you are confusing your logical fallacies. At worst, the part you quoted would be argument from collective experience which I have never heard of as being a logical fallacy. PS : Was this a troll? By the definition that anything that runs contrary to Erik's opinion is a troll, yes it was a troll. Someday hopefully you discover the wide world of alternative opinions. Ooo, nice ad hominem! Erik -- +---+ Erik de Castro Lopo +---+ C++ : You won't live long enough to learn it all from experience. -- Peter Miller (author of Aegis) -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Question on ownerships and permissions of subversion respositories
On Sat Apr 29, Michael Lake wrote: Hi all I have a question on ownerships and permissions of subversion respositories. /var/lib# chown -R root:src svnrepos/ Actually I'm also planning to have the repository browsable by Trac (an SCM system) so I'll probably need it also readable by www-data. What does one do there if one wants users to be able to have file:// access locally and apache to have read access? Mike -- Mike Lake Caver, Linux enthusiast and interested in anything technical. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality (and software testing)
[EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 With regards to last night's Slug meeting and using automated testing, I think everyone agrees that writing (and using) test cases produces higher quality code with less bugs. My point is that higher quality output doesn't come for free, it requires effort and that usually means someone has to pay for it. If you are writing code because you want to write code and you want to produce the best result that you can then sure you are going to want to put the extra time into it and do work you can be proud of. Use whatever techinique you think will work, regression tests will help, good design documentation will help, feedback from users helps too. On the other hand, if you want to get a job done, get paid and get out of there (and that's the way 90% of business works, sorry to say) then the fact is that corners get cut and the end result is not high quality. With Open Source style programming, it might eventually end up as high quality software once enough people have got interested enough to bash it into shape (and often after the second or third re-write from scratch) but it very rarely starts out that way. In the early days of a project it is hard enough to get enough time and effort in to make it work at all (even as a buggy proof of concept) let alone produce a high quality masterpiece. That's why I follow this simple and easy to remember rules: Make it RUN; Make it RIGHT; Make it FAST; and Make it NICE. Every software developer test their codes whether the testing procedures are formalized or not. I take it to mean that we are debating whether to formalize or not the testing procedures every time. I don't agree with this dogmatic approach to formalizing test procedures every time. Circumstances do vary and good software developers adjust to varying circumstances. Hope this helps. O Plameras -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Serious Printer Question -- Samsung CLP-550N
Have you looked at Brother, when i was doing my TCO calculations Brother was also very cheap and they have very good Linux support. On Sat, 2006-04-29 at 18:55 +1000, [EMAIL PROTECTED] wrote: I'm thinking about buying one of these printers, they are available under $800 and they support Postscript and PCL5c as well as the undocumented (i.e. useless) Samsung Printer Language. They also have a rather low cost-per-page when you add up the consumable costs (note that the KYOCERA FS-C5020N has a lower cost per page but you have to print about 4 pages to break even, when you consider the different up-front costs, I doubt I'll print that many). Regards Richard Neal Real Men don't make backups. They upload it via ftp and let the world mirror it. -- Linus Torvalds -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Question on ownerships and permissions of subversion respositories
On Sat, Apr 29, 2006 at 10:22:50PM +1000, Michael Lake wrote: I have a question on ownerships and permissions of subversion respositories. I'll be using svn+ssh. I have set up repos on an open/free OS for use with svn+ssh. svn+ssh means that repo access is protected by normal Unix user/group controls. Our users already have ssh access to the machine holding the repos, so no special ssh stuff is required - just their normal login keys. Each distinct work group has all members belonging to the same /etc/group, and I set the group of all repo files to match. This just means you need to put all workers for the same project into the same /etc/group - but they're probably already there. If the project needs different group membership, it needs a different group, and a different repo. My repos are all fsfs. I put wrappers around all svn binaries to set umask to 002, and do chown -R svn:group repo chmod -R ug+rwX,o= repo Because I don't use web or other access to these repos, I don't really care whether the ownership is svn or something else, as all access to this stuff happens because of g+rw. If you do want public read access, you might change repo modes to ug+rwX,o+rX. A not so good alternative might be to chown it to the user they all get to read as (say something www-ish) - in which case you might want u+rX,u-w. (My repos are not public, so I haven't tried either of these out.) -- Christopher Vance -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Hardware Question
Hi all I am building a Production FC4 box with Scalix (www.scalix.com) for use in a small site (12 users). The site is My church and I don't have Bucketloads of $$ to spend but I have two available boxes to build on. 1Ghz Celeron with 1.5Gb SDRAM (3 slots) 2.8Ghz P4 with 1Gb SDRAM (2 Slots) Now the question is this. For running Scalix (Mail server) and general Samba stuff. Am I better to lean towards the Faster Processor with Less RAM or do I go with the More RAM but slower Processor? From what I understand I cannot buy SDRAM in sticks bigger than 512Mb so each Box is Maxxed out. Can anyone give me an educated opinion on what the best to do is? Ta - Kevin Fitzgerald TCG TECHNOLOGIES Skype: kevtcg MSN: [EMAIL PROTECTED] Mobile: 0412 404 002 Web: www.tcgtech.com.au -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Re: Paying Money for Quality (and software testing)
On Sat, Apr 29, 2006 at 01:49:43PM +1000, [EMAIL PROTECTED] wrote: With regards to last night's Slug meeting and using automated testing, I think everyone agrees that writing (and using) test cases produces higher quality code with less bugs. My point is that higher quality output doesn't come for free, it requires effort and that usually means someone has to pay for it. cost(Debugging without tests) cost(debugging with tests) + cost(tests) Why? Because your tests tell you where the problem is, usually. If your test suite isn't comprehensive enough to localise the problem sufficiently, then the first part of your debugging routine is to add tests to your test suite that localise the problem. This is time that you would spend localising the problem in an ad-hoc fashion anyway, and if you encode it in your test suite, then it's time spent only once (writing the tests) instead of spending the time localising the problem to the same unit over and over and over again. This single point is enough to win me over to test-first development. It has cut the amount of time I spend debugging any non-trivial program by a huge amount. Then there's all of the other factors, which improve my job satisfaction -- and a happy programmer is a productive programmer, which improves productivity as well. But they're far more intangible. - Matt -- Windows is too dangerous to be left to Windows admins. -- James Riden, ASR -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Permission problems on linode (Ubuntu 5.10)
Asking here because I know some other people use Linode. (For people who don't know, a Linode is a UML image provided by linode.com. One important thing to note is that using a Linode means I am using a kernel build of theirs, not a distro provided kernel. Nor do I have opportunity to build my own.) I have a Linode running Ubuntu 5.10, which I upgraded from their Ubuntu 4.10 image (via Hoary/5.04). It has ubuntu-minimal and ubuntu-standard installed, so udev is installed as a dependency. The other day I changed from using a 2.4 kernel to their 2.6 kernel: # uname -a Linux [hostname] 2.6.15-linode16 #1 Wed Jan 4 17:30:25 EST 2006 i686 GNU/Linux Since this upgrade, I am not getting the correct permissions on some /dev nodes, in particular: # ls -l /dev/null crw-rw 1 root root 1, 3 2006-04-30 11:54 /dev/null This is causing a fair number of problems on my system: many processes expect to be able to redirect or write to /dev/null. I understand udev is meant to take care of making the device nodes for me and setting the correct permissions, what do I need to change about it? -Mary -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Serious Printer Question -- Samsung CLP-550N
On Sunday 30 April 2006 10:00, [EMAIL PROTECTED] wrote: Have you looked at Brother, when i was doing my TCO calculations Brother was also very cheap and they have very good Linux support. On Sat, 2006-04-29 at 18:55 +1000, [EMAIL PROTECTED] wrote: I'm thinking about buying one of these printers, they are available under $800 and they support Postscript and PCL5c as well as the undocumented (i.e. useless) Samsung Printer Language. They also have a rather low cost-per-page when you add up the consumable costs (note that the KYOCERA FS-C5020N has a lower cost per page but you have to print about 4 pages to break even, when you consider the different up-front costs, I doubt I'll print that many). My mate bought a Brother. Nice in every respect except the inks in a colour page crack and flake off if the page is folded. ie print a colour A4. Fold it in 3 to mail. The inks along the folds flake off. Is this a generic problem, or a Brother one? James -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Hardware Question
IMO, for what you are talking about, both machines are over-spec'd. Email is not time critical and 12 users and Samba will hardly cause any box to crack a sweat - we're not talking Windows here. If I had this situation I would use the 1GHz box, pull out 1Gb of RAM and sell it to recoup some of the costs. Win - win all round. On Sun, April 30, 2006 12:02, Kevin Fitzgerald wrote: Hi all I am building a Production FC4 box with Scalix (www.scalix.com) for use in a small site (12 users). The site is My church and I don't have Bucketloads of $$ to spend but I have two available boxes to build on. 1Ghz Celeron with 1.5Gb SDRAM (3 slots) 2.8Ghz P4 with 1Gb SDRAM (2 Slots) Now the question is this. For running Scalix (Mail server) and general Samba stuff. Am I better to lean towards the Faster Processor with Less RAM or do I go with the More RAM but slower Processor? From what I understand I cannot buy SDRAM in sticks bigger than 512Mb so each Box is Maxxed out. Can anyone give me an educated opinion on what the best to do is? Ta - Kevin Fitzgerald TCG TECHNOLOGIES Skype: kevtcg MSN: [EMAIL PROTECTED] Mobile: 0412 404 002 Web: www.tcgtech.com.au -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html -- Howard LANNet Computing Associates http://lannet.com.au When you want a computer system that works, just choose Linux; When you want a computer system that works, just, choose Microsoft. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Permission problems on linode (Ubuntu 5.10)
On 4/30/06, Mary Gardiner [EMAIL PROTECTED] wrote: # ls -l /dev/null crw-rw 1 root root 1, 3 2006-04-30 11:54 /dev/null It needs the rw other permission set, how you do that for a udev device not sure (as I dont suspect doing chmod 666 /dev/null will do it. (althougth this is how you would do it for a normal file). Sorry not upto scratch on udev.. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Mounting Flash Drive
Would some kind slugger please tell me what I have to type into the terminal to mount my flash drive in Fedora 4? John. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Serious Printer Question -- Samsung CLP-550N
On Sunday 30 April 2006 12:19, [EMAIL PROTECTED] wrote: On Sunday 30 April 2006 10:00, [EMAIL PROTECTED] wrote: Have you looked at Brother, when i was doing my TCO calculations Brother was also very cheap and they have very good Linux support. On Sat, 2006-04-29 at 18:55 +1000, [EMAIL PROTECTED] wrote: I'm thinking about buying one of these printers, they are available under $800 and they support Postscript and PCL5c as well as the undocumented (i.e. useless) Samsung Printer Language. They also have a rather low cost-per-page when you add up the consumable costs (note that the KYOCERA FS-C5020N has a lower cost per page but you have to print about 4 pages to break even, when you consider the different up-front costs, I doubt I'll print that many). My mate bought a Brother. Nice in every respect except the inks in a colour page crack and flake off if the page is folded. ie print a colour A4. Fold it in 3 to mail. The inks along the folds flake off. Is this a generic problem, or a Brother one? I'd say it is a Brother one as I have never encountered this problem with the Samsung. The only problem I've encountered is printing on card stock where the toner simply flakes off. This can be overcome by setting the paper to card. I notice the Duplex unit is optional (extra cost) on the Brother whereas the Samsung is built in which is an extra bonus I think. -- Regards, Graham Smith -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Mounting Flash Drive
Is it a usb device? if so the following would do.On mine in the command prompt I used to type:mount -t msdos /dev/sda1 /mnt/usb-diskmsdos (is the file system, this can be any file system like ext2, ext3 etc.) /dev/sda1 (usb disk's use the scsi driver. like the hard drive its sd instead of hd) /mnt/usb-disk (is the location to mount it)Pete.john gibbons [EMAIL PROTECTED] wrote: Would some kind slugger please tell me what I have to type into the terminal to mount my flash drive in Fedora 4?John.-- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html On Yahoo!7 Answers: Real people ask and answer questions on any topic. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Paying Money for Quality
On Saturday 29 April 2006 13:49, [EMAIL PROTECTED] wrote: snipped I've found a good practical example to demonstrate this... snipped The answer is that these people (and there are a lot of them) don't want to pay that bit extra for a quality printer. Adding Postscript to a printer is very much like adding an automated regression test suit to your software. If you buy the same model of printer with Postscript, you don't get any extra printer, no extra dots per inch, no extra speed, no bigger paper stack, etc. You get exactly the same physical printer as the model without postscript. What you do get is better compatibility, better long term reliability (because you know it will work with anything you give it), less time wasted debugging problems, less headache, less heartache, less hair loss. ... but you gotta pay money for that ... snipped Printers on the brain? You need to report to room 101 for retraining. Consumer electronics are bought by consumers not customers. They have no needs beyond compatibility with the current Microsoft OS release. If it doesn't work for them, it must be the consumers fault (and they should be reported as a possible subversive too). I don't think your example is really all that correct, so it's probably not the best one to illustrate what you were trying to discuss about coding practices. Cheers from Minitrue, Malcolm V. -- To save a single life is better than to build a seven story pagoda. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Mounting Flash Drive
On Sunday 30 April 2006 14:45, Peter Hodder [EMAIL PROTECTED] wrote: Is it a usb device? if so the following would do. On mine in the command prompt I used to type: mount -t msdos /dev/sda1 /mnt/usb-disk msdos (is the file system, this can be any file system like ext2, ext3 etc.) /dev/sda1 (usb disk's use the scsi driver. like the hard drive its sd instead of hd) /mnt/usb-disk (is the location to mount it) Wouldn't it be better to use 'vfat' instead of 'msdos'? I don't think msdos supports long file names, for instance. Another option would be to use 'auto' to have the filesystem automatically recognised, which is useful if you also use other filesystems on USB media. -- Sridhar Dhanapalan [Yama | http://www.pclinuxonline.com/] {GnuPG/OpenPGP: http://dhanapalan.webhop.net/yama.asc 0x049D38B4 : A7A9 8A02 78CB AB1B FCE4 EEC6 2DD9 249B 049D 38B4} Windows XP is not an operating system. It is a windowing system that sits atop an operating system much as KDE or Gnome sit atop Linux. - Robert X. Cringely, 2003-01-16 pgp3jiipzDj5H.pgp Description: PGP signature -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Mounting Flash Drive
john gibbons wrote: Would some kind slugger please tell me what I have to type into the terminal to mount my flash drive in Fedora 4? John. I though core 4 would just detect in. Have you tried plugging it in and then looking with your favourite file browser in /media/usbdisk Does this not work?? Sorry to ask the obvious first. Otherwise, Plug in your flash drive. Open a terminal Type dmesg without the quotes and press enter About 6 or so lines up from the bottom, there should be a line starting: usb-storage: device scan complete Following that line, you should have a line that says SCSI device sd?: (the ? should be a or b or c or so on) At the end of that line should be a capacity in MB such as mine that says (65 MB), that should be the capacity of your flashdrive. Type mount -t vfat /dev/sd? /media/usbdisk The sd? obviously needs to match what you read in the dmesg output. Hope this helps. D. -- Never ascribe to malice that which may adequately be explained by incompetence. - Napoleon Bonaparte -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Firefox 1.5 (FC5) and Plugins
How do you install plugins in FF 1.5 (FC5) I can find no reference to plugins in any of the menus, nor in the preferences. I went to a site that required JRE so I downloaded it using the offered plugin finder screen (jre-1_5_0_06-linux-i586.rpm.bin), ran the install, and - nothing... When I go back to the site it still says I don't have JRE installed. Nowhere in FF can I find any reference to determine what plugins are installed. This is stupid, and frustrating. -- Howard LANNet Computing Associates http://lannet.com.au When you want a computer system that works, just choose Linux; When you want a computer system that works, just, choose Microsoft. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Firefox 1.5 (FC5) and Plugins
OK, stupid me - read ALL of the installation instructions - including the symlink instructions. Sorry for the unnecessary noise. On Sun, April 30, 2006 15:11, Howard Lowndes wrote: How do you install plugins in FF 1.5 (FC5) I can find no reference to plugins in any of the menus, nor in the preferences. I went to a site that required JRE so I downloaded it using the offered plugin finder screen (jre-1_5_0_06-linux-i586.rpm.bin), ran the install, and - nothing... When I go back to the site it still says I don't have JRE installed. Nowhere in FF can I find any reference to determine what plugins are installed. This is stupid, and frustrating. -- Howard LANNet Computing Associates http://lannet.com.au When you want a computer system that works, just choose Linux; When you want a computer system that works, just, choose Microsoft. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Firefox 1.5 (FC5) and Plugins
Howard Lowndes wrote: How do you install plugins in FF 1.5 (FC5) I can find no reference to plugins in any of the menus, nor in the preferences. I went to a site that required JRE so I downloaded it using the offered plugin finder screen (jre-1_5_0_06-linux-i586.rpm.bin), ran the install, and - nothing... When I go back to the site it still says I don't have JRE installed. Nowhere in FF can I find any reference to determine what plugins are installed. This is stupid, and frustrating. The plugin filename is /usr/java/jre1.5.0_06/i386/ns7/libjavaplugin_oji.so. You have to make a symbolic link (ls -s) into /usr/lib/mozilla/plugins. Copying will not work. Hope this help. O Plameras -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Permission problems on linode (Ubuntu 5.10)
quote who=Mary Gardiner The other day I changed from using a 2.4 kernel to their 2.6 kernel: Can you try their 2.6.16 kernel and report back? (That's what I'm running, it's finally TLS-happy too.) - Jeff -- linux.conf.au 2007: Sydney, Australia http://lca2007.linux.org.au/ Instead you're doing circle jerks with the Care Bears of Censorship. - Siduri on Slashdot -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html