Slashback: Debianism, Nukes, Discretion 161
Lots of pretty pictures. An unnamed reader wrote to say "that notes and pictures from the Debian Conference are now available." There are lots of cool (somewhat KDE-centric, go figure) shots from the recent LinuxTag in an article at dot.kde.org, too ;)
Adobe: an especially thick, recalcitrant material. aicra writes: "according to an email sent to the list:
Cute nuclear material shipment, has all shots, answers to "Lucky." We miss it very much. Please call. Random Walk writes: "The Russian memo, and the email exchange between Russian and U.S. principals involved in the problem of nuclear material lost in a MS database, is available online from the CDI. I found the following sentence particularly funny: 'Relatively poor quality of Microsoft SQL Server has created very serious problems in development and implementation of CMAS.' There is a very detailed discussion of the problem(s) with MS SQL Servers 6.5 and 7.0. The Russian report also has interesting words about the 'common fault failure' problem and the need to be more careful when selecting software for critical systems."KIllustrator is now known as 'Kontour' (after Adobe claimed that KIllustrator was too close to Adobe Illustrator, in case you haven't followed the news).This obviously breaks translations... I have kept the name killustrator.po[t] for the message files, so this remains. But the appearances of the word KIllustrator have been replaced.
This delays the release by about 2 days -- so that there is time to get the new messages translated for rc1 if you're fast (and for the final release otherwise)."
Take this exchange for what you will. As michael posted the other day though, the issue isn't necessarily all (or even primarily) Microsoft. The importance of robust software doing the dirty work doesn't get much clearer though.
Absolutely, positively, undeniably, unmitigatedly maybe. Later. Last week, we relayed a report that Psion was getting out of the consumer products market. An Anonymous Coward writes now, however, that "according to this article at Psion Place: Peter Bancroft, a senior spokesman for Psion, released a follow-up statement saying that 'There will be more Psion consumer products in the future.' According to Bancroft Psion is merely 'suspending' their plans to develop a consumer Bluetooth device."
The same statment, though, indicates that if Psion does eventually develop consumer products again (with hints about Bluetooth), it may make them in combination with a partner, or sell off the rights entirely.
KIllustrator... (Score:1)
Re:How often the MSSQL errors occurred... (Score:1)
While no one competent in the corporate IT field uses MS SQL Server for anything important (that's what Oracle is for), there's plenty of incompetents in the world, and the amount of real money that could simply go astray due to the cumulative errors in gimcrack MS systems is scary.
MORAL: Don't use MS software for anything that matters. Preferably, don't use it at all. If you don't like Open Source, then your still better off with a non-MS proprietary vendor...
Re:Blamethrowing (Score:2)
-Paul Komarek
Psion will return? Yeah, sure. (Score:1)
Yep, and when Apple killed the Newton, they announced that they would return to the handheld market in 1999.
We're still waiting.....
Geoff
Re:It clearly is an MS SQL Server Error (Score:3)
--
"Shared" Source != open or free (Score:2)
clean room (Score:2)
Re:Korel Draw (Score:1)
I think "PageMaKer" or "FrameMaKer" would be too klose for komfort.
Regards, Ralph.
Re:Psion and the consumer market (Score:2)
It does.
The EPOC OS may be the greatest thing since sliced bread, but if you can't download the email addresses you entered in your PDA right to your mail program with ease, it's useless.
It can, so it isn't.
A CD-ROM containing the PsiWin [psion.com] PC-Psion sync software is included with every Psion computer.
Regards, Ralph.
Re:Gods (Score:2)
Re:Psion and the consumer market (Score:2)
I can't give Palm too many props in the software area, though - though their desktop software upgrades are free, it took nearly a year after Win2K was released to have a non-beta quality USB driver. That's way too long.
- -Josh Turiel
Re:Psion and the consumer market (Score:2)
Murky screen: Yes, I've seen the contrast setting and fiddled with it. It's still pretty crummy. However, two counterpoints to that: first off, the Revo/Revo Plus is the entry-level Psion, and competes closest to Palms. And the screen is much crisper on Palms, even the lower-end ones. I know the 5mx has a much nicer screen (with backlighting), but it's also a lot bigger and heavier.
"Just deal with it" doesn't cut it. Hinges can be engineered for a lot less flex pretty easily - I think they sacrificed that in the design for the telescoping back and port covering when folded. The Revo normally costs some serious $ - it shouldn't feel flimsy.
I haven't seen the Linux software for it yet, but I'll go hunt and see how good it is. I do not have a problem with Outlook (or Windows, for that matter - other than the fact that I need to use it at work whatsoever, but that's another story!) - and I might mention that _nothing_ else I own has problems in this regard. The PC doesn't lose track of the Palm, Jornada, or Blackberry. However, the older EPOC Connect software Diamond included didn't work at all (never could find the Revo), and the new (beta, but the best they offer) version from Psion still drops the connection.
As for cost - I realize the Psions are positioned higher up. But that's too small a market. If you want to compete for profits with Palm and Microsoft, you need to have software developers. To have more developers, you need more users (the bulk of them go where the money is). To have more users, you need market share - and to get market share, you need PDA's that appear to the consumer, not just the geek.
And no, I didn't buy it used. I bought it new from Outpost.com, who is dumping the Diamond Mako (a Revo Plus with a different badge) along with Sparco and a few other vendors, since Diamond/Sonicblue is getting out of the Psion business. It was shrinkwrapped and everything.
Obviously, I haven't figured out everything it can do - four days is just enough to scratch the surface. I have downloaded quite a few apps to tinker with and learn, and don't get me wrong. I do consider it a neat device, I like it, and I'm glad I picked it up.
I'm also glad I didn't pay list price for it. Because after four days, it is a cool device, but I can tell readily why Psion has bombed so utterly in the general consumer market. Because when I take off my geek hat and try and look at it like the average person who might go to a Staples or something for a PDA, it does not measure up. People like you and I, however sophisticated we may be, do not make a marketplace.
- -Josh Turiel
Psion and the consumer market (Score:5)
So after using it for a full weekend +, I can see the usefulness of it to some people but I also can see why Psion's made pretty much zero dent in the consumer PDA market as a whole. I'd played with some other models before, but the Revo Plus is the first one I've owned.
Good points to the Psion:
-When folded, the shell is pretty solid. It's the only plastic PDA I've ever felt comfortable with in a pants pocket (the converse, though, is it's a little tall for shirt pockets).
-The weight is comfortable to carry.
-Battery life is excellent, and there's clear battery life indicators.
-It's pretty quick - switching applications is fast once you get used to the clunky way to do it (or download a nice free task switcher).
-EPOC apps seem to generally be pretty compact. With 16MB of RAM you can cram a lot o stuff into it.
-Licensing Opera: Good Move!
-Like CE, you can browse the palmtop's filesystem when it's docked from your PC. It makes transfers pretty easy.
Unfortunately, it's balanced by things that truly suck:
-The screen (the Revo is non-backlit) can be murky.
-When open, it feels pretty flimsy and flexes in my hand. It's much nicer on a table than in the hand.
- The connectivity software included with the Diamond version is horrid (an older EPOC Connect version). Downloading the current version from Psion (and patching it to the latest fixes) makes it usable, but it still has an alarming habit of disconnecting itself from my Windows laptop with no warning.
-The CopyAnywhere software for Psion/Windows clipboard sharing sucks hard.
-Syncing with Outlook (which I hate, but it is the standard for us) results in Outlook popping up to ask permission to share the e-mail addresses. Outlook thinks PsiWin's a damned macro virus! My Palm and Blackberry don't have this problem - they actually integrate with Outlook well.
-Syncing with my Mac at home - the Palm does it for free (with a free download or software on the disc nowadays, I believe). Buy a Mac, and the software to sync a Palm is already on the hard drive. Psion charges $65 for it, and all it can do is file backups.
-On a related note, all you can easily share from Outlook is the Contacts and Calendar files (which are probably the strongest of the built-in Psion apps, IMO). And the notes from Contacts don't transfer. Also, the Notes section of Outlook doesn't map to anything at all (no conduit), so I had to export to Notepad and copy the
Basically, I like the unit, it's useful, rugged (when closed), has a nice geek factor to it, and I don't regret buying it at all. That said, I think the average consumer would be far better served with a Palm (or maybe CE), and here's why:
-Palms are far cheaper on average. Easier impulse buy.
-A Palm has much more seamless connectivity with the two main desktop platforms out there (Mac and Windows), and it's all included for free with the organizers. Linux support is pretty good, too.
-Palms are smaller, lighter, and more easily pass the shirt pocket test. They also feel more solid when being used. Pocket PC's usually feel like tanks (and excepting the iPaq, are usually built like them) - they're even more rugged than the Psion is.
Re:veKtor (Score:1)
--
Re:Psion and the consumer market (Score:3)
Yes, the Revo has a lower res and smaller screen. It makes some apps difficult.
The task switcher is odd, but switching tasks is easier if you just click on the icon on the launcher bar, it does the same thing.
I have a 5mx and a Series 7. The 5mx is a little slower (for most things, but a lot for CPU intensive items). It works well, but the 7 has a 640x480 STN color screen and a larger keyboard. The 7 is a little bulky to carry, but the 5 isn't as nice to use for long periods.
All in all, a 5 is a killer item, and if you have a GSM phone (I bought one just for it), it integrates really well. Phoneman can manage your phone contacts and send SMS and ringtones.
The spreadsheet is killer, and Jotter (notepad) works really well for making little freeform databases (like conversion tables).
It has real SQL query support for database files, via the OPL interface. Also, the contacts/database functionality is pretty sophisticated.
If you can afford a Series 7, get it, it supports CF type 2 and PCMCIA type 2. It's fast and can hold 32MB of RAM. The enterprise version has an NE2000 ethernet adaptor built in. If you do a lot of writing, the keyboard on the 7 is awesome.
If you can't afford a 7, the 5mx is your next best choice. It has all the same features, yet it's B&W and smaller (about 40%). It uses 2 AA batteries and runs about 20hours without the backlight on. It's keyboard is functional, but many common characters require FN-key combinations.
The Revo is an excellent competitor to PIMs, but is too small for a lot of keyboard use. The screen is 480x240, so it has the display area of a 5 with the toolbars turned on. On the upside it has really tiny pixels which make everything look extra sharp. It also comes with a docking cradle and Phoneman builtin.
The EPOC series of devices are lightyears ahead of any other PIM/PDA OS and application suite. EPOC is by far the most functional OS for a PDA. It doesn't have handwriting recognition, but rather uses a keyboard. The pen is like a touch screen mouse. It is well laid out and intuitive to use. It makes very good use of screen real estate and file storage. The 16MB that comes in most is storage AND memory to run, which shows just how efficient EPOC is.
EPOC is the OS of the future for every handheld consumer device, Nokia switched to it for the new communicator series, and the new phones are awesome.
That's Funny... (Score:1)
-Waldo
Re:That's Funny... (Score:1)
-Waldo
Re:Kontour (Score:1)
Have fun with the confusions.
Far from a satisfactory result (Score:1)
This isn't a satisfactory result, except that I'm happy the developers aren't risking their careers/incomes by putting themselves in line of legal fire. This doesn't mean I'm happy with Adobe for laying claim to a word that used to be part of the English language. Far from it.
--
Dieresis (Score:1)
Well, I was wrong (Score:2)
I said here [slashdot.org], and I quote:
Guess I won't be taking any more crap from Microsoft defenders the next time I assume that Microsoft can't cut it - the one time I trust Microsoft not to have screwed things up, they go and let me down.
Yep, that last laugh is pretty satisfying, at least until my major city disappears in a ball of nuclear fire :)
Tax dollars wasted (Score:1)
I think directly flaming MS over their product quality is pointless. There was once, a long time ago, a point to this. But nowdays, Microsoft's track record and reputation are well-established and known to everyone, so responsibility falls not upon Microsoft, but upon whoever makes the decision to use their products. It's kinda like if a monkey shoots someone: you don't blame the monkey, you blame whoever let the monkey have a gun.
Usually there's a reason for companies to use Microsoft products such as MS SQL. The most common reasons are:
- They don't know better
- A MSCE (either actually certified, or a wannabe in the same mindset)
employee or consultant, as part of the lock-in an job security strategy,
has set them up to do that. Some PHB in IS has probably set a policy
that these are the tools people must use.
My normal reaction to this is: Do nothing, because it isn't my problem. The company will be natually punished/rewarded for its own ignorance and gullibility. If I were a stockholder, I might wish that the people responsible were fired and replaced someone more competent or honest. But realistically, I know that whatever infintesimal fraction of a share I own indirectly through some mutual fund, is utterly powerless.Here we have the government using Microsoft products, and using them for important things. I don't see this as being any different than the classical stories about $900 hammers and toilet seats. The "it's not my problem" approach that applies 99% of the time, doesn't apply in this instance since it partly my tax dollars that are being misappropriated, either through fraud or ignorance. But the "infintesimal fraction of a share" argument still applies. So what the fuck can be done?
---
Re:How often the MSSQL errors occurred... (Score:2)
It depends on the job at hand. If a Fisher-Price toy plastic hammer does a poor job of driving actual metal nails into wood, that doesn't really mean it's a poor product. The hammer's job is to entertain children, not drive nails.
MSSQL is no different. The problem isn't that it fails one in a thousand, the real problem is that someone responsible for working with nukes, thinks that 999/1000 is good enough.
---
Re:Blamethrowing (Score:1)
While I agree with the point you are trying to make, MS SQL 6.5 is still the only commercial db I've seen that lets you wipe out a whole table with a single mouse click and gives no warning or confirm dialog. I've seen a reasonably competent DBA do exactly that.
If you are trying to figure out how to do this, think bidirectional replication and a disaster recovery situation that involves installling a replacement server. (Which is something that does come up a bit in maintaining critical systems).
Re:The K Syndrome... (Score:2)
And yet you are still reading
--
Re:Blamethrowing (Score:2)
Safer than they were anyway - safer to the point of (almost
Re:Blamethrowing (Score:4)
Designing safe kid's toys is not easy! Note that I am a qualified mechanical engineer, with experience of meeting safety requirements for a pedal car, so this isn't just me talking crap.
At the same time, I really think that kids these days are just too safe - all they seem to learn is how to be good little consumers...
Re:Shared Source (Score:2)
This might be a cultural thing, but where I'm from, if something doesn't show up where it should be, we describe it as being "lost" (until it is "found")
--
Matt
Re:Blamethrowing (Score:4)
Jeezus, do they still _make_ pedal carz?
I thought they had been outlawed in principle - for safety reasons.
I remember those well. They were a _great_ way to do all kinds of foot injuries, from pinching them in the moving parts to scraping them on the ground. There were, of course, the whole gamut of collision class problems (other kids had pedal carz, too) and then there were the running-off-the -road style problems (just put one over a curb - with or without help from another vehicle - a rollover is a distinct possibility, with pinched fingers pretty much guaranteed).
But the _real_ way to really really kill yourself was to get a bigger kid (the bigger the better - kinda like the big-block V8 engines that were being mounted in real cars at the time) to shove you up to an honest 15MPH or so. At that kind of speed, anything could happen - except that you couldn't get yer feet anywhere _near_ the pedals, so braking was pretty much out of the question.
It sure was a good thing we didn't live near any good, steep hills. The kids who lived near hills had _scars_.
All I see in toy stores anymore are electric jobs with a top speed of about 2 mi/hr, and reduction gearing so strong that there's no possibility of pushing them any faster.
DON'T GET ME WRONG! Pedal cars taught me vital lessons: that I wasn't indestructable, and that moving concrete sidewalk can _hurt_ raw skin. Both these proved invaluable as soon as I learned to ride a bicycle....
You can't tell me any amount of engineering in the world can make those things "safe".
Lego? Fsck that - I was known to eat _staples_!
Re:Long/Lat Coordinates (Correction: Tennessee) (Score:1)
Long/Lat Coordinates (Score:4)
Slashdot's traveling, I guess.
It clearly is an MS SQL Server Error (Score:5)
Additional tests were developed in order to detect a systematic appearance of this SQL Server 6.5 software flaw. These tests were constructed in a such way that a presence of a flaw in the SQL Server 6.5 has been proved for sure. Microsoft did confirm a presence of this software flaw. This flaw has been filed by Microsoft as flaw No. SRX000403600845.
In a course of transition of KI-MACS application software core from version 6.5 to 7.0 and intensive testing of the modified KI-MACS, at the end of July, 2000 one more flaw has been detected in the SQL Server 7.0. This flaw has been recognized by the Microsoft and filed as flaw No. SRX000727603512. Principal modifications introduced by the Microsoft in SQL Server 7.0 which are relating to the data security and data protection in a Database, and presence of the said flaw, are considered as creating a direct threat to data security and data protection in the SQL Server 7.0 Databases.
One in a thousand (Score:2)
Fear the Uncertainty (Score:2)
The system with bugs apparently uses Microsoft Access (therefore using a
Microsoft operating system). Developed at Oak Ridge, it involves a PC
which monitors containers which rest on weight sensors. Some
installations also use electronically-readable tags, heat sensors or
radiation sensors. Russian seals were used, not the
electronically-monitored Oak Ridge seals. Announced plans were for 15
Russian facilities to be using the system.
http://www.sandia.gov/LabNews/LN02-16-96/la1995.h
http://www.y12.doe.gov/orsens/pubs/inmm99report-r
http://www.y12.doe.gov/orsens/pubs/esarda-vane.pd
http://rmtc.ippe.obninsk.ru/seminars/eng/tri2/NMC
slow link)
The U.S. NRC had a different problem with a Microsoft database in 2000.
The Nuclear Materials Management and Safeguards System was first
implemented with Microsoft FoxPro, which "the computer industry no
longer supports".
http://www.nrc.gov/NRC/COMMISSION/SECYS/2000-0163
The OPTI-RP software for selecting radiation protection options uses
Microsoft Access.
http://ean.cepn.asso.fr/pdf/EAN-2.pdf
Microsoft Access is used for a Singapore inventory of Air Emissions,
Watewater, Solid Waste and Toxic waste.
http://www.dnauk.co.uk/env-pro.htm#it12
Southern Africa's 20 Megawatt pool type reactor SAFARI-1 is automated
with a Windows NT system, including material tracking. Also: "...we are
looking at remote reporting and further user interfaces at the homes of
senior plant personnel, for faster response time to incidents requiring
operator intervention"
http://www.adroit.co.za/news/briefs.html#ITEM27
The Australian Radiation Incident Register uses Microsoft Access.
http://www.arps.org.au/ARPS25abstracts.PDF (350K on slow link)
The 2000 nuclear Information System on Occupational Exposure (ISOE) 1
and 2 database uses Microsoft Access.
http://www.nea.fr/html/rp/reports/2000/nea2728.pd
RTFA - see SRX000403600845 & SRX000727603512 (Score:1)
Microbob
Re:It clearly is an MS SQL Server Error (Score:1)
__
Country Club Road, not Way, but yeah ;) (Score:2)
There are blackberries growing just off the backyard here, and the area is not yet converted into 100% suburban wasteland. (Though this development is certainly doing its part
Cheers,
timothy
Re:Blamethrowing (Score:2)
It is inconsequential that hundreds of thousands of other sites seem to run SQL Server 6.5, 7.0, and 2000 just fine.
Well, first of all, the bugs discovered by the Kurchatov Institute does actually exists, and acknowledged as such by Microsoft (see the mail from the MS engineer). Note the bugs are bugs within MS SQL Server, not in the implementation of the DB. So 'Oh yes, it's all SQL Servers fault.'
I don't care to check whether the problems is fixed by now, but I guess they are, though probably not for MS SQL 6.5. (read the text)
So every MS SQL server not patched, has the potential to be bitten by these bugs.
A nice quote from the text:
stupid advises as change of data formats to avoid MS SQL Server SYSTEM
problems.
Is it MS style of conducting business? If so, we have to be prepared to
deal with permanent troubles for the rest of our life.
Re:Simply amazing... (Score:1)
----
Re:GIMP can't replace PS until CMYK patents expire (Score:1)
----
Re:Kontour? (Score:1)
----
Re:GIMP can't replace PS until CMYK patents expire (Score:2)
And I can do a CMYK conversion in a non-Photoshop program and bring it into Photoshop with all four channels intact and correct. Did it all the time when I last worked in-house for a printer, batch converting several hundred RGB images into CMYK.
----
Re:Shared Source (Score:2)
Have you ever used an RDBMS in anger? Your statement is analogous to saying 'if you delete a file from your file system, no data is actually lost - it's just not showing up where it should be'. That's true, of course. If you go and inspect the inode, you'll find the data is still there - at least until it gets overwritten by something else. But it's not good to you because, without doing esoteric things in the bottom of the file system, you can't get at it.
The SELECT statement is the principal way of getting data out of an RDBMS. If it doesn't work, that RDBMS is broke. If it doesn't work and it doesn't barf, that RDBMS is not just broke, it's dangerous. I've spent twenty minutes this morning writing a heads-up to all my customers who use MS SQL Server, advising them of the problem.
Re:Blamethrowing (Score:2)
And you can prove this how?
Remember the nuclear materials accounting system was in use at Los Alamos for several years and they never detected thec error; the Russians were only able to detect the error because they had duplicate paper records. Do you have a papertrail (a real one, on real flattened dead trees) covering all your transactions? If not
Because, face it, man. If you are using SQL Server 6.5 (or have done in the past), and you use (or any software you use uses) ORDER BY, you have lost data. Like the guys at LANL, you just don't know it yet.
Spot vs. process colour (Score:2)
>Sure, it might work "as good" on your screen
>(your screen can't show CMYK
>correctly anyway), but when you are going to be
>printing things on a press, you
>need to have support for real Pantone colors.
>Adobe continues to have a
>monopoly in this area.
The normal concept of Pantone spot colour is quite different from that of CMYK process colour. The latter requires (in the instance of computer modeling) a method to convert from RGB display on-screen to CMYK output (note that GhostScript does this, but is buggy and imprecise
Pantone is a commercial standard which anyone can make use of by _licensing_ it from the company---NeXT did this for NeXTstep, so drawing and other DTP programs were cheaper for it, since no extra license was necessary.
However, one can do spot colours without such a license---just create a channel/separation/plate and tell the printer, ``this plate should be blue, i.e., Pantone 301'', but it won't necessarily look like that on the screen (a common trick is to say cyan==blue, magenta==red, and such like, to get spot colour blending which the typical dtp apps don't handle). This is way cool if one has a RIP or other trapping/ink manipulation tool which allows ink-substitution.
Sadly, spot colours are misunderstood even in the graphic design industry (one Quark book author stated he didn't understand where the Pantone color numbers came from, they're from the ink formulation, so Pantone 301 is three parts cyan, 0 magenta, one yellow, no black)
The problem with the GIMP for spot colours is there's AFAICT no multi-channel support, with an understanding for colour as applied pigment and which models the difference between additive and subtractive colour.
William
--
Lettering Art in Modern Use
Re:Shared Source (Score:2)
Errhm, not quite, and in the following lines yourself are actually saying why...:
> If you go and inspect the inode, you'll find the data is still there - at least until it gets overwritten by something else.
Whereas, with this database bug, the data won't get overwritten (unless you DELETE it first...)
> But it's not good to you because, without doing esoteric things in the bottom of the file system, you can't get at it.
The real problem is that you won't expect a simple SELECT ... ORDER BY statement to fail, and fubar happens if your application program acts on the result it gets from the statement (for updating other tables, displaying reports to the users, sounding (or not sounding...) an alarm about missing nukes, etc.).
Re:Blamethrowing (Score:1)
SELECT count(*) FROM table WHERE {conditions};
SELECT count(*) FROM table WHERE {conditions} ORDER BY {ordering};
and look at the difference in numbers?
Does SQL Server work with SELECT count(*)'s and ORDER BY's, but not with SELECT {data} ... ORDER BY's?
Re:Blamethrowing (Score:1)
Indeed not. That's like citing Lego as an example of engineering excellence.
Re:Shared Source (Score:1)
Re:Shared Source (Score:1)
by pnatural on 08:56 PM July 16th, 2001 EST (#96)
(User #59329 Info)
Ever heard of SELECT INTO? A bug there would cause data to be lost. Or how about a stored proc that performs an INSERT or UPDATE based on the content of a SELECT? Add DTS into the mix, you have a million and one ways a bug in SELECT would cause data loss.
[ Reply to This | Parent ]
Re:Shared Source (Score:2)
by mikeee on 09:06 PM July 16th, 2001 EST (#107)
(User #137160 Info)
>no data is actually being lost - its just not showing up where it should be.
You're saying that the data isn't lost, you just can't find it?
Hello!
And when it comes to nuclear security, I'll take transparency over features any day; put the damn records in a flat file if you have to.
[ Reply to This | Parent ]
Re:Shared Source (Score:1)
by Tuross (root@[127.0.0.1]) on 09:17 PM July 16th, 2001 EST (#113)
(User #18533 Info) http://www.goldweb.com.au/~matt/
Also, if it is just a bug in SELECT statements, no data is actually being lost - its just not showing up where it should be.
This might be a cultural thing, but where I'm from, if something doesn't show up where it should be, we describe it as being "lost" (until it is "found")
--
Matt
"Hope you don't think I'm rude - FUCK YOU!!"
- - - - -
Re:Blamethrowing (Score:2)
- - - - -
Re:How often the MSSQL errors occurred... (Score:2)
- - - - -
Re:Blamethrowing (Score:2)
- - - - -
Re:veKtor (Score:1)
Re:Shared Source (Score:1)
Re:Lost Data (Score:2)
I don't think reliable means what you think it means. From reading the article it looks like proper care means adding extra checks to detect database corruption. It's a continuing struggle because each release has different bugs.
Re:Adobe Email Addresses? (Score:2)
Re:Precisely my point. (Score:3)
In addition, there were flaws in KI-MACS that were not. But saying Microsoft is blameless is a lie.
Re:Gods (Score:1)
It did rock
Re:Psion and the consumer market (Score:2)
The technology the programmer sees... the operating system, the API is irrelevent. The technology the end user sees is king.
The most important technology for PDAs is the ability to sync with their desktop mail & contact information. This is a technology that Psion doesn't have. The EPOC OS may be the greatest thing since sliced bread, but if you can't download the email addresses you entered in your PDA right to your mail program with ease, it's useless.
Also, on a more mundane note, the reason the first poster got the Psion for $100 was because of something called *a sale* (on a discontinued item, no less). You might look into the concept - in particular, it doesn't contract the fact that Psions are generally more expensive than Palms.
So that is why ... (Score:1)
"Our hardware runs better without Windows" [www.hack.gr]
?Re:Lost Data (Score:1)
http://slashdot.org/features/01/06/10/0034241.shtm l [slashdot.org]
Hint: how come no one bothered to comment in that article ?
Just for the record, I think that it's better to loose data with a GPL'd database than with a Microsoft one.
Re:Lost Data (Score:1)
Korel Draw (Score:2)
Phillip.
Re:Psion and the consumer market (Score:1)
However, should point out that the problem with Outlook thinking it is a macro virus is also a Microsoft one - it's part of their ridiculous SP2 security features. It assumes anything trying to access your address book is a virus.
Re:Blamethrowing (Score:1)
Re:Shared Source (Score:4)
Re:Lost Data (Score:3)
Indeed. However, most people rely on the fact that the database server itself won't lend a willing hand in the destruction of your data by feeding you incorrect information from SELECT commands.
MS SQL Server is alive and kicking and pervasive throughout enterprise computing. With proper care, it can be as reliable and strong as Oracle and leagues stronger than MySQL or PostgreSQL.
Reliable is not random data loss. If SELECTs and ORDER BY ops are not returning all the data they should, there is something seriously wrong, and the best DBA in the world probably can't do a damn thing about it. One would expect that some esoteric features might be buggy, but one that is as ubiquitous as SELECT? RDBMS's live and die by the data integrity they provide -- I wouldn't run MySQL for a mission critical system that could not ever lose or corrupt data because it was never made for that. MS SQL and Oracle are both designed to do this, and if MS SQL can't deliver, I'd call that a pretty big reliability problem.
Re:mysql loses data all the time. (Score:3)
Never use MySQL to store anything important. Use PostgreSQL, InterBase, or some other Open Source DBMS which passes the ACID test.
Check here [openacs.org] for a discussion of the problems with MySQL.
The difference, though, is that MySQL does not pretend to be a real DBMS. It's problems are well known and documented. Even Las Alamos couldn't mistake it for one. MS SQL, on the other hand, is marketed as a enterprise level RDBMS. Consequently, its users have an expectation of a resonable level of quality.
In short, it is acceptable for MySQL to lose data, it is not acceptable for PostgreSQL, Oracle, Interbase, MS SQL, or any other database in their league to behave in a similar manner.
Debian Conference notes updated (Score:2)
Just thought to note that I've done a major update to my Debian Conference page [www.iki.fi], somebody of course posted it just before I had my report ready. :-)
AJT
Re:Shared Source (Score:4)
You're saying that the data isn't lost, you just can't find it?
Hello!
And when it comes to nuclear security, I'll take transparency over features any day; put the damn records in a flat file if you have to.
Dateline: Moscow (Score:2)
From: Mikrosoft
Re: New databaseski
Date: 11/1/1989
Comrades,
We offer you some of our new software which is as strong and reliable as your venerable Berlin Wall. May the Communist Party, founded on the ideals of Lenin and supported by MS SQL Server(tm), reign for another century.
Love,
Bill
======================================
Re:Ohh, Wait A Minute (Score:2)
"The Software product may contain support for programs written in Java. Java technology is not fault tolerant and is not designed, manufactured, or intended for use or resale as online control equipment in hazardous environments requiring fail-safe performance, such as in the operation of nuclear facilities, aircraft navigation or communications systems, air traffic control, direct life support machines, or weapons systems, in which the failure of Java technology could lead directly to death, personal injury, or severe physical or environmental damage."
Re:Blamethrowing (Score:2)
How the hell can you blame Microsoft for your stupidity on not being able to handle the most common input device?
It's a sad day on slashdot when people are sinking to this level.
I've installed several instances of multiple SQL Server 7 databases, usually as a temp-store in a DMZ with automagic replication to a main server, and I've never seen so much as a bit of data disappear.
Admitedly the code was written by professional coders and not hacked together in 20 minutes, but the data was & is still there.
Please people. I remember when MS started pushing SQL server, and people were pleased it was based on the (rather good) Sybase codebase, so at least keep your flames subjective rather than whining.
Let those who manages real, large systems make intelligent comment on this. I don't, but I know people who do, and I really am curious to hear about real world problems.
What I don't want to hear about is some'l33t MySQL server with 10 records in it doing a better job from someone barely out of school.
Re:Blamethrowing (Score:3)
But by this logic Unix has an inherent failure in letting any idiot do rm -r when they've just backed up one level from a symlink, and trashed a vital directory structure! My point was that blaming a standard human-computer interface allowing you to delete a database hardly constitutes MS evil.
I don't even dare apply Windows NT or MSSQL7 service packs to it
Try it, with SQL server it usually works. If it doesn't then revert to your backup.
That said, I do use it myself at work for the website and the online ordering for non-critical data such as pricing, and minimal customer data. If information is lost, it's not important, because it will not affect operations at the company one bit
I think you may find "online ordering" is fairly mission critical for those websites who rely on the Internet as a shop-front, and thousands of those rely on MS SQL Server. I rarely recommend MS products, and even rarer when they are used as components of web servers (I'm a security consultant after all), but I think the product is being flamed far more than it deserves here.
I'd deploy Oracle on Solaris 8 days a week if I could, but it's expensive, full of magic (none of which is in manuals, it's smoke & mirrors DBAing), and is only cost-effective for real cash-generating business entities. If you want a reasonably straightforward system with cheap support costs (in terms of manpower), then there are far worse things than deploying SQL server.
Shared Source (Score:2)
Then they could fix the bug in their database that the Russians found.
For all the MS-bashing on this story, the bug seems a bit esoteric. All software has bugs. I doubt they would be better off with an open source database given the mission-criticalness of the application. The best open-source database is still incomplete compared to any commercial RDBMS.
Also, if it is just a bug in SELECT statements, no data is actually being lost - its just not showing up where it should be.
How often the MSSQL errors occurred... (Score:5)
Lost Data (Score:2)
Re:Lost Data (Score:2)
Almost always (note almost) the problem with pretty much all sql servers is due to the admin not knowing enough about it to run it properly. I've seen some really crappy Oracle servers running on 32 proc Sun systems along with some really kick ass Oracle servers on a 4 proc Sun system. The DBA makes or breaks sql systems these days a whole lot more than the sql server software itself.
Re:Blamethrowing (Score:3)
mysql>show databases
LittleData
BigData
mysql
mysql>drop database BigData;
ok, 0 rows affected
mysql>show databases
LittleData
mysql
mysql>
Damn, looks like I wiped a whole database(not just a table) with a single command with no warning or confirm dialog. The point is any inept administrator can lose data. You mention replication, but what good does replication do in this instance? The database gets dropped from your replicated DB as well. Time to find the tapes....
Enigma
Re:How often the MSSQL errors occurred... (Score:2)
[sigh]
and MS wants us to trust them with the .NET
Check out the Vinny the Vampire [eplugz.com] comic strip
Looking at these pictures... (Score:2)
The K Syndrome... (Score:4)
KIllustrator is now known as 'Kontour' (after Adobe claimed that KIllustrator ...)
We need to ban the letter 'K' from all Linux software. The misspelled words are Killing me.
What's next? kmdrTaco?
Re:Kontour (Score:2)
Re:Kontour (Score:2)
-mdek.net [mdek.net]
Re:GIMP can't replace PS until CMYK patents expire (Score:3)
Sure, it might work "as good" on your screen (your screen can't show CMYK correctly anyway), but when you are going to be printing things on a press, you need to have support for real Pantone colors. Adobe continues to have a monopoly in this area.
-
Ban the letter "k" (Score:2)
By banning the letter "k", distributions would have to refer to their "kernel" package by its real name---Linux! It almost makes too much sense!
As a corollary, maybe they would then call their distributions something besides "Linux". Perhaps "operating system".
veKtor (Score:2)
Unsettling MOTD at my ISP.
Re:It clearly is an MS SQL Server Error (Score:2)
~
mysql loses data all the time. (Score:2)
Looks like the middle of a Golf Course (Score:2)
Skynet, er, SQL Server (Score:2)
Obviously they made the wrong choice. Now we can truly blame MS for the end of the world! Who knows, maybe SQL server is really Skynet in disguise... (see movie Terminator)
Re:Russkies aren't used to the free market, it see (Score:2)
Having worked with many a Russian programmer, I can attest to their competence as programmer. They do like oderly processes and document their systems to a fault, but the software works as specified (even .0 versions) and if you find a bug, they fix correctly and quickly. Most of the times, they find the problem before we report it.
Found the problem (Score:2)
I think it's time they stopped using MS-Access. Obviously the transactions aren't working right. Nor is the 'repair db' button. No matter how many times that was clicked, the data's still missing. Port attempts to FoxPro have yielded similar problems, and Excel now refuses to load on their server.
Nuclear Material (Score:2)
Apparantly the security there was so low that he just had to put it asides and take it with him home. check out [spiegel.de] the german magazine "spiegel online".
You may need babelfish [altavista.com]
Governmental buck passing (Score:2)
I find it hard to believe that MS releases a commercial server product that actually loses data as a result of it's normal running. As users of MySQL, the slashdot staff should be aware of the precautions necessary with any database use -- even fault-impervious Open Source Tee Emm databases -- to prevent data loss and allow recovery. I place the blame for the missing nuclear material in the hands of the government adminsistrators, who were obviously too lazy to do their jobs.
I will agree that choosing the correct RDBMS solution is crucial. Smart companies use Oracle, not MS._ _____
____________________________________________
MS SQL (Score:2)
D/\ Gooberguy
Re:Governmental buck passing (Score:2)
Well duh. They work for the Gov't. Being lazy IS thier job!
Carl G. Jung
--
Re:Ohh, Wait A Minute (Score:2)
Carl G. Jung
--
Re:Adobe Email Addresses? (Score:3)
Carl G. Jung
--
Re:Blamethrowing (Score:2)
2) The Russians were given the program by the Americans. Offhand I don't remember who developed it, but it was a custom job.