Stop laughing. We do it, and it works well. TortoiseSVN makes it really easy.
It doesn't take much technical sophistication to handle "update" and "commit", and that's 95% of the operations on this sort of repository. Very little branching, some use of logs...but really, what people need is a place to put documents that fires off commit emails and where it's possible to get a log or pull an old version if necessary.
As far as the sales guys are concerned, it's a lot like a network share, except that they can still access their local working copy when they're on a plane or at a customer site.
Even ignoring viruses/worms altogether, it's not a good idea for users to be exchanging.DOC,.XLS, and.PPT files through email. People do this for two reasons:
Exchanging finished documents for reading. PDF is better:
It can reproduce the results exactly.
It doesn't include Word's "change tracking" information which can cause embarrassing leaks.
It's a standard with many interoperable implementations.
Exchanging in-progress documents for revision. At least for stuff limited to your company, a version control server (like Subversion with friendly TortoiseSVN clients) is better:
Doesn't cause email storage to grow enormously. Instead, a server actually meant for this kind of thing stores only deltas. And only one copy of each document - on most mailservers, the disk space consumed by an attachment is proportional to the number of recipients.
Lets you easily find the latest version of a document. ("Did he send me another copy after this? I'm not sure.")
Lets you easily retrieve any previous version, see changes/authors/checkin comments. (I don't trust Word's built-in change tracking, and you shouldn't either. Its security model is flawed, and I don't think it's reliable to begin with.)
Supports locking/unlocking documents to prevent conflicting changes.
With some setup, supports diffing and merging office documents. You can maintain branches!
Supports searching - where I work, we've plugged in swish-e for full-text searching over our documentation repository.
I wish my company would just block all.DOC and.XLS files sent from one employee to another. It'd force them to use the documentation repository and save us all a tremendous amount of pain trying to dig through email for the right version of some Product Requirements Document. It'd also stop the whining from people complaining about hitting their email storage limits all the time.
I now have some no-name-brand CDs burned in 1998 that are still good.
Probably using the original dyes, then? According to the article, they are most likely to fail in 2008:
The first organic dyes, designed by Taiyo Yuden, were Cyanine-based and, under normal conditions, had a shelf life of around ten years; simply, that was simply unacceptable for archive discs.
These people are talking about serious long-term archiving, not "it worked for this one guy for eight years".
I haven't been using DVD-R long enough to comment on how long they'll last.
No one has successfully used them as long as these people are talking about; they haven't existed that long. The lifespan claims are made from an understanding of chemistry (theory) and accelerated aging techniques (experiment).
This is a particularly blatant and well-documented example, but it's not surprising. Verizon regularly lies to consumers, actively or by omission.
When I signed up, I had no credit history, so they charged me a large deposit which was to be returned after a year. When I called after over a year asking where my check was, they told me that I had to request the deposit to be returned. Who has ever heard of such a thing? Why didn't they mention this when I started the account? They were simply hoping I would forget that I'd paid the deposit or wouldn't be willing to fight them for it. How many deposits have they just kept in this way? Or put another way, how much of other people's money became Verizon's because of deception? How much money did they steal?
But what can you do about it? There's no accountability. "George" and "Andrea" are either absolutely incompetent or dishonest, but they don't even tell you their full names. You can't link the voices on the phone to actual people. Even if you could, there's no channel to complain about them. And there's certainly no way to link the absence of an action to a specific person, so there's certainly no way to hold them accountable for not sending my check. And unfortunately, you can't just switch to a more honest phone company, because I don't believe such a creature exists.
I think the most that can be done is to take them to small claims court each time. If you go through all the work to do so, you'll almost certainly win. But they're betting most people don't have the time to fight them, and...well, they're right.
I was about to make the same post before seeing this lurking down at Score:0. SMTP does guarantee delivery or notification of failure! It provides a clear statement of whose responsibility each message is at each point, and how notification works when delivery fails or is refused. This excuse is wrong.
I just read the first paper. It indeed defines a new number system, but his description is wrong:
[The paper] describes how to divide by zero consistently in a non-trivial way. This shows that division by zero is no longer an error. Amongst other things, the paper explains why the standard model of arithmetic is not valid.
Unfortunately, that explanation seems to have been replaced by gibberish in the copy I just downloaded. Check it out:
Unfortunately, IEEE floating-point arithmetic is not a valid model of arithmetic either. We cannot accept an arithmetic in
which a number is not equal to itself (NaN =? NaN ), or in which there are three kinds of numbers: plain numbers, silent
numbers, and signalling numbers; because, on writing such a number down, in daily discourse, we can not always
distinguish which kind of number it is and, even if we adopt some notational convention to make the distinction clear, we
cannot know how the signalling numbers are to be used in the absence of having the whole program and computer that
computed them available. So whilst IEEE floating-point arithmetic is an improvement on real arithmetic, in so far as it is
total, not partial, both arithmetics are invalid models of arithmetic.
So basically, the two NaNs have subtle semantics (much like his nullity) and don't have a catchy name or reuse a symbol that already means the golden ratio, therefore they're broken.
[NaN's] semantics are not defined, except by a long list of special cases in the IEEE standard.
In other words, they are defined, but he doesn't like the definition.
So a function with some nullity arguments may perform arbitrary processing on them,
because they are just numbers. A database record with value nullity is not set to any real value. A time stamp with value
nullity is not set to any real time.
Right. Now my airplane won't drop out of the sky, because the thrust calculation that used nullity as an input produced nullity as an output, in a way completely different from the one that produced NaN from NaN before. This new name and slightly different semantics magically mean the right amount of fuel will go into the engine.
Every calculus student knows the answer to such questions as "what is the limit of 1/x as x approaches 0 from the right" (positive infinity) and "what is the limit of 1/x as x approaches 0 from the left" (negative infinity). And more usefully, derivatives are defined as "the limit of (f(x+h)-f(x))/h as h approaches 0 from the right". Integrals are defined through limits as well. So properly phrased (through limits) questions like this are the very foundation of calculus, and well-understood.
But "1/0" alone? What does that even mean? There's no answer. It's a stupid question, too oversimplified to distinguish between the possible answers. The first thing he wrote - "infinity = 1/0" - was already wrong.
If he made his arguments to his peers instead of schoolchildren, they'd shoot him down, and rightfully so.
Furthermore, saying that computers cannot divide by zero shows a ridiculous lack of common sense. They can do anything we design them to do. Many computer number systems (notably including IEEE 754) have a special value NaN (not a number) that is similar to his nullity concept, except that it's not arrogantly proclaimed as revolutionary or a solution to every problem. Generally, asking a question such as "1/0" indicates a serious logic error. Imagine that airplane needs to calculate the proper elevator trim. Oh, great, the answer is nullity. What does that mean? How should it move the elevators? Giving this failure condition a new name doesn't change the fact that the airplane's still going to drop out of the sky.
Interesting! I'll have to look it over more later.
For comparison, I've put the latest (not yet deployed) version of our offline checkpoint process here. (It's a NetVault backup script; pre locks and does the checkpoint, post touches a file signalling success to our monitoring and releases the lock). It's a procedure outlined by Perforce, though they didn't mention error handling...
It means that (for example) while the hour-long checkpointing pre-backup process happens every night, you can't do any write operations.
Let me be a little more specific: while the hour-long checkpointing process is happening, you can't even open files for edit. In addition to having really course locking, Perforce has more write operations than most version control systems. Subversion's CVS-style working copy means the only write operations are commits and revpropsets.
Good idea, building on a closed-source SCMS that's (barely!) a mid-level player in the market. I can understand not wanting ClearCase, but what's wrong with CVS or Subversion?
I use both Subversion and Perforce. There's one major feature still lacking from Subversion: merge tracking. There's work underway to design, implement, and document this feature, but it's not done yet. This is a huge deal for anyone with lots of branches.
Not that it's all roses with Perforce. My impression is that it doesn't scale very well. Most operations simply lock the entire database. I think it's a reader/writer lock, but it means that (for example) while the hour-long checkpointing pre-backup process happens every night, you can't do any write operations. (And there's a way to do an offline checkpoint, but it's not documented or supported, and is difficult to get right, with bad consequences if you don't.)
When I said they didn't care about "us" I meant the users of RedHat, but I guess everyone assumed I was speaking for the entire Open Source and Free Software community.
RedHat's past, present, and future contributions (including to core software such as gcc, the Linux kernel, and glibc) benefit everyone who uses Linux, and nearly everyone who uses any open source software at all. That includes you, even if you are a freeloader, and even if you no longer use RedHat Linux or Fedora.
So they are helping you now. Do they give a fuck about you? Okay, maybe not. I certainly don't; you're a jerk. But that's a useless question - you are taking advantage of their work anyway. Maybe even my work, though you probably don't know it.
[The convenant RedHat broke is] the "you scratch my back, I'll scratch yours" bit. Except in this case it's "I'll act like Microsoft, and you'll eat it up because we sell Linux!"
There is no such convenant...but if there were, they have not broken it. Formally, the statement "False -> True" is true. You have never scratched their back, and they continue to scratch yours. This makes you a hypocrite, and RedHat a great member of the community. Your "they act like Microsoft" lies do not change that.
Am I not allowed to have an opinion if I didn't write emacs?
You can have an opinion, but I'm done listening to it, and it appears the moderators are as well.
Trolling is when you say something you don't believe to elicit a desired response.
Right. Isn't that what you were doing? Or do you really not see the obvious inconsistency between RedHat's extensive open source contributions and the idea that they don't give a fuck about the community? Is there a third choice? I don't see it...
Oh, and this is the list you were looking for.
Close, but that's not the one. I've seen a much more extensive list, with some details about each their involvement in each project, explicitly broken down into the categories I mentioned. I think it was on the Fedora wiki.
In the interest of full disclosure, you should mention that you have Foe'd me when you ask someone to mod me down.... No wonder you foe'd me, I must have had endless reasons to cuss you out for stupidity in the past.
No, I foed you immediately after asking for you to be modded down. I don't ever recall hearing of you before; your post just pissed me off.
Re:Missing a Chapter
on
Fedora Linux
·
· Score: 4, Insightful
Wake me up when the book has a chapter entitled "How Red Hat Software broke its covenant with the linux community" talking about how the supported stable free version of the software disappeared, to be replaced by a beta test program for RHEL.
There never was such a convenant, and RedHat releases all the source packages as required by the GPL. This means you can download a RHEL-based stable, free distribution recompiled by a third party in the form of CentOS. My company tends to buy RHEL where we want paid support or where we have to buy an operating system with hardware (Dell offers Windows or RHEL at the same price). We use CentOS elsewhere. It's the same software, and it's free and legal.
Many of us, myself included, were formerly RedHat users but we learned that they don't give a fuck about us when they shifted from free RedHat to Fedora, and haven't looked back since.
For a company that doesn't give a fuck about open source, they sure hire a lot of people to develop it. I can't find it now, but somewhere there's a webpage with a list of all the open source projects RedHat has developed, maintains, and contributes to. It's amazing how much they've given back to the community, and how many idiots like you there are who are deliberately blind to it.
Can someone please mod the parent down as a troll? And someone else please post a link to the page I can't find?
Re:package manager need tons of work
on
Fedora Linux
·
· Score: 5, Informative
One problem yum has had is that it wants to check the network for updates before every operation. This has improved recently... It still needs to re-read the data, which takes longer than it should, but only has to call out to the network if something is likely to be different, which makes a *huge* difference when you're installing individual packages or querying it with search or info.
Reading and re-reading the data should be quicker now, too.
The repository data is stored in a giant XML file which is incredibly slow to parse. Back in the day, it would read this file in every time you ran yum. Last year [1] they added a SQLite cache, so this step could be skipped if the data hadn't changed.
Relatively recently, they added a separate yum-metadata-parser written in C that dramatically reduces the time the parse step takes. Take these changes together and what used to take 45.5 seconds every time you ran yum now takes 7.5 seconds only if the data have changed. [2]
It sounds like they've done as much as they can without changing the transferred data to be an indexed binary format (with the associated forward/backward compatibility complexity).
(I'm not running Fedora Core 6, so I'm not sure if this change made it in.)
[1] - Looks like yum 2.3.1 introduced the cache, around March 2005.
[2] - See this message introducing it around May 2006 sometime after yum 2.6.1.
Put simply, this was WRONG. The kid deserves to be arrested, NOT electrocuted.
They couldn't arrest him because he didn't commit a crime. Here's what the article says:
Tabatabainejad, 23, was shocked Tuesday night after arguing with a campus police officer who was conducting a routine check of student IDs at the University of California, Los Angeles, Powell Library computer lab.
Campus police say he refused to show his student ID and refused to leave the building when asked.
Note the questions it doesn't answer - what business police had checking IDs in this computer lab or forcing people to leave a public building, what law he broke.
You can't arrest someone for a made-up crime...but apparently you can pump over 50 kV through his body five times in rapid succession. Strange world...
The sourceforge job listing in question says this:
We're continuing to grow here at SourceForge.net! We have recently opened a position for a Senior Java Developer, and are accepting applicants. Think you got the stuff?
The position is in our Fremont, California headquarters, but we are open to applicants in the United States who are willing to travel frequently. So if you want to work in a flexible, creative environment and know how to do things well, this position might just be right for you.
Let's assume for a second that I'm a superstar programmer. Why would I want this job? It says nothing! Frankly, sourceforge's website does not appear technically sophisticated, and the existing stuff is all in PHP. What would I be doing? Why would it interest me? Even super-secretive, super-prestigous gives way more information than this. If you want super-star programmers, you're competing with google. You don't have their prestige, so you need to do something better than them. This is it.
You are in good company, though. Much of the time I see places asking for superstar programmers without giving the slightest idea why they're needed or what they'd be doing. It's not a successful strategy...which is probably for the best, because a superstar programmer would be bored out of his (or her) skull if he actually got the job.
I used examples relevant today. If you want older ones, they're essentially unchanged:
How long would it take to send a 1 MB file over a 1.544 Mbps DS1?
How many times per second does the clock cycle on my 4.77 MHz IBM XT?
How long would it take it transfer the contents of said diskette over a 2400baud modem?
Probably less long than you'd think, if any such modem ever existed. 2400 bps modems were 600 baud.
(clockspeed has always used the standard definitions AFAIK. 1Khz = 1000Hz, etc.)
It's khz, small k. Now I understand why you don't see the problem. It causes headaches for people familiar with the SI prefixes outside of computing.
Different, but consistent.
That phrase is nonsense. If they were consistent, they'd be the same, by definition. You previously said that the computer industry was consistent. I showed that only one of three situations within the computer industry used the definition you gave, and you agreed. It's not consistent, and there are many situations where conversions between these three situations causes confusion.
I see that you've gotten a few answers already, so I'll only post mine where they differ. (In other words, my complaints. I'm generally happy with OS X, though - enough so that I ordered a new MacBook Pro when I saw the announcement this morning.)
3) How stable is it. Macs are traditionally easy to use, but as I've owned dozens of Macs (and used to sell them too) I can attest that they were not the most stable machines out there (up until the first OSX spin). But browsing the knowledge bases and user forums (the BEST place for info) I see lots of issues.
Pretty good. If your hardware is working, the kernel will be solid.
Finder will be pretty unresponsive if a network drive has stalled - pretty normal for a Unix system, since non-blocking IO is unfortunately not supported on regular files. (The only system I've seen get around this is KDE with its ioslave design.) You can force quit it and it'll come back.
Some stuff takes up annoying amounts of CPU time. Safari will suck down all my CPU (and thus battery life) when on some pages, presumably because of a Flash or JavaScript interaction. Mail + Spotlight really get lost in some massive "Public Folders" hierarchies on my employer's Exchange server. I really wish they had a way of honoring IMAP subscriptions or telling it not to locally cache certain paths. You'll want to run with MenuMeters always visible so you can spot this before discovering that your battery is dead only an hour into the flight.
4) How much Free software is available? Can GNU/Open/Free programs be compiled easily and natively? I'd think because it's more consistent than the hundreds of Linux distros, this would be true...
Fairly easily. The compiler and linker arguments around dynamic libraries are a little weird, but the fink package system (like *BSD's ports) will generally patch things for you, and libtool-based stuff won't have a problem. I basically just type fink install foo and whatever I want shows up.
5) How solid is the workmanship. Hey, I get mocked at work because of my Dell, but it was cheap and it's fast. That's usually all I need. My Thinkpad is better built, but the $600 price difference was not worth it. What makes the Mac worth the extra $1K?
I don't think their laptops are really any better-constructed than the next guy's. I find all laptops to be pretty fragile - I've had way more problems with them than with desktops. I'm not normally an extended warranty guy, but I'd recommend it on any laptop. AppleCare is quite nice. You can also buy it at any time up to the last day the normal one-year warranty expires.
7) How does the two-finger trackpad stack up against real buttons? I.e., it's software to emulate two physical buttons. I've not used it before. Any drawbacks?
I didn't know about this feature, but on a later PowerBook, I used the two-finger scrolling. It works wonderfully.
Well, it was different but atleast consistent until the marketing department of the harddisk manufacturers got involved and screwed it all up for the rest of us. And for the most part, they are the only ones still at it. My "512MB" stick of memory contains 536,870,912 bytes of memory, not 512,000,000.
Ignoring for a second that "different" means "inconsistent", you're still wrong - it was never consistent even within the computer industry. Pop quiz:
How long would it take to transfer a 1 GB file over a 512 kbps link? [*]
How many bytes on a 1.44 MB floppy?
How many times per second does the clock cycle on my 1 GHz Athlon processor?
Use the new definitions. "1 GiB" is much clearer, and eventually people will know that "1 GB" means something else in new software. This sort of switchover is not unprecedented - IIRC astronomy switched to use the same notation for the polarization of light as engineering. Papers around the time would carefully note which definition they used. Ones well before or well after are obvious. More complicated for the computer industry, but not impossible.
[*] - Massive extra credit for correctly considering the protocol overhead for...we'll say FTP over PPPoA VC/Mux over ADSL...but that's not the point of the question.
With the expectation settings thing, I'd have to say "It depends." I would give high marks to someone who had read the situation clearly and was able to come up with the "right" answer without needing to be told all the details.
I see what you're saying, but on the other hand, I'm thinking of a story I read last night. Umm, there's a telling of it here. Short version: Enrico Fermi met some "great" generals and asked what it was that made them great. They decided that it was winning five consecutive battles. And...about 3% of generals were great. Well, that's the same number (.5^5) he would have expected by random chance if all battles were completely even, and all generals were alike.
My point is: maybe the people who gave your interviewer what he was were looking for just got lucky. They might have used the same process (guessing or assuming) as the ones who were wrong. I don't see anything in your story that makes me think that there was enough information to figure out what your interviewer wanted, if only they were smarter. If you ask people to read your mind, some people will get the right answer...but they're still no good at mind-reading.
Though if I were the interviewee, I probably would have done the same thing you did. It seems safer to initially guess real-world rather than contrived. If I say "I wouldn't solve this problem; I'd change it" and they're looking for that, I'm done. If they're not, (1) they'll justify it with a real-world detail and we'll continue, (2) they'll say "this is a contrived interview question; just answer it" and we'll continue, or (3) they'll count it against me, which is just as well because I'll count it against them, too.
Asking questions is good...depending on the question. I've had a few people actually ask "what are you looking for?" The first time, I (stupidly) gave away the whole point of the question, so it became worthless. Other interviewers said that this person was very good at manipulating an interview, but not necessary at the job. Since then, I'll just say "for you to solve this like you would in the real world...except where I specifically tell you otherwise".
The point of an interview is to figure out how well the person will fit and how well they'll do the job. Clever hacks and workarounds are nice, but only when they are more efficient and effective than something not as clever would be. It seems like a lot of geek types forget that - it isn't about showing off.
Half of that is the interviewer's fault, for not setting expectations. Sometimes you want to see how they'd solve something in the real-world way, and sometimes you want to see how smart/resourceful they are by using contrived situations. "No, you can't ask your coworkers for help; they've all died in a bizarre skiing accident. And you must use this expensive, broken, completely inappropriate product as the basis of your code, because the CTO Has Spoken.[*]" If you made it clear which is which, some people would probably do much better.
I'm starting to wonder if I'm guilty of this myself - my favorite interview question is oversimplified and inherently inefficient. I basically expect them to accept the stupid specification as is (some whining is okay) yet implement it in a real-world way - I emulate "man", "gcc", and "gdb" for them. I try to make that clear, but maybe I could do better...
[*] - Sadly, this second part actually happens. I've seen "you must use Oracle RAC", and I've heard of much worse examples.
If Windows eats your partition table, you can be pretty sure that either you are having problems with LBA modes and translation, you botched the partitioning in the first place, or you botched the part of partitioning that you do in windows.
I am sure that none of those things happened. Read the KnowledgeBase article I linked to. I've also added to my advogato post the actual partition layout in question.
The whole point is that Windows doesn't know it's formatted as another filesystem. It doesn't know any filesystems other than FAT/FAT32/NTFS. All it sees is a partition in no known format, and assumes it is just random unformatted bits. It isn't Windows' fault that you didn't check your partition table (always, always triple check that before formatting drives!!!).
Are you being deliberately dense? My partition table was right; I've pasted the actual contents here. Windows doesn't need to understand ext3 to know that partition type ID 0x83 means "something I shouldn't mess with".
I believe the note is slightly inaccurate. It should say "If you have a partition before the partition you are installing Windows to, that is formatted anything other than FAT/FAT32/NTFS, or unformatted, it must be reformatted along with the Windows partition."
That seems more in line with the KB article I linked to, but it's a huge difference. This badly violates the principle of "the system should not touch an existing partition of a foreign type without the user's permission". And "reformat unformatted partition?" doesn't count - it should be smart enough to know the answer is no, but at least the question should be "reformat as NTFS extended partition 2 of size XXX that's marked as ext3?"
I'm almost positive it asks twice, but I can't refresh my memory without formatting the first partition to get to the second prompt, which would have to be one of my working partitions I have now.;-)
That's an important observation. Even if you are right about the two prompts (and maybe we have slightly different situations; it sounds like both of your partitions are tagged as NTFS), it's still badly broken. If it's already destroyed your data by the time you know anything is wrong, there's a huge problem.
Oh, by the way:
PartitionMagic sure comes in handy and I wish there was a full featured GPL alternative. Anyway, that's all way offtopic.
GNU parted doesn't speak NTFS, but I believe it can do all the Partition Magicy things to a Linux system.
Slashdot Mirror
It doesn't take much technical sophistication to handle "update" and "commit", and that's 95% of the operations on this sort of repository. Very little branching, some use of logs...but really, what people need is a place to put documents that fires off commit emails and where it's possible to get a log or pull an old version if necessary.
As far as the sales guys are concerned, it's a lot like a network share, except that they can still access their local working copy when they're on a plane or at a customer site.
- Exchanging finished documents for reading. PDF is better:
- It can reproduce the results exactly.
- It doesn't include Word's "change tracking" information which can cause embarrassing leaks.
- It's a standard with many interoperable implementations.
- Exchanging in-progress documents for revision. At least for stuff limited to your company, a version control server (like Subversion with friendly TortoiseSVN clients) is better:
- Doesn't cause email storage to grow enormously. Instead, a server actually meant for this kind of thing stores only deltas. And only one copy of each document - on most mailservers, the disk space consumed by an attachment is proportional to the number of recipients.
- Lets you easily find the latest version of a document. ("Did he send me another copy after this? I'm not sure.")
- Lets you easily retrieve any previous version, see changes/authors/checkin comments. (I don't trust Word's built-in change tracking, and you shouldn't either. Its security model is flawed, and I don't think it's reliable to begin with.)
- Supports locking/unlocking documents to prevent conflicting changes.
- With some setup, supports diffing and merging office documents. You can maintain branches!
- Supports searching - where I work, we've plugged in swish-e for full-text searching over our documentation repository.
I wish my company would just block allThat's cool. I don't know if they changed their practices (mine was a year ago IIRC) or if you just won the lottery.
Probably using the original dyes, then? According to the article, they are most likely to fail in 2008:
These people are talking about serious long-term archiving, not "it worked for this one guy for eight years".
No one has successfully used them as long as these people are talking about; they haven't existed that long. The lifespan claims are made from an understanding of chemistry (theory) and accelerated aging techniques (experiment).
This is a particularly blatant and well-documented example, but it's not surprising. Verizon regularly lies to consumers, actively or by omission.
When I signed up, I had no credit history, so they charged me a large deposit which was to be returned after a year. When I called after over a year asking where my check was, they told me that I had to request the deposit to be returned. Who has ever heard of such a thing? Why didn't they mention this when I started the account? They were simply hoping I would forget that I'd paid the deposit or wouldn't be willing to fight them for it. How many deposits have they just kept in this way? Or put another way, how much of other people's money became Verizon's because of deception? How much money did they steal?
But what can you do about it? There's no accountability. "George" and "Andrea" are either absolutely incompetent or dishonest, but they don't even tell you their full names. You can't link the voices on the phone to actual people. Even if you could, there's no channel to complain about them. And there's certainly no way to link the absence of an action to a specific person, so there's certainly no way to hold them accountable for not sending my check. And unfortunately, you can't just switch to a more honest phone company, because I don't believe such a creature exists.
I think the most that can be done is to take them to small claims court each time. If you go through all the work to do so, you'll almost certainly win. But they're betting most people don't have the time to fight them, and...well, they're right.
I was about to make the same post before seeing this lurking down at Score:0. SMTP does guarantee delivery or notification of failure! It provides a clear statement of whose responsibility each message is at each point, and how notification works when delivery fails or is refused. This excuse is wrong.
Unfortunately, that explanation seems to have been replaced by gibberish in the copy I just downloaded. Check it out:
So basically, the two NaNs have subtle semantics (much like his nullity) and don't have a catchy name or reuse a symbol that already means the golden ratio, therefore they're broken. In other words, they are defined, but he doesn't like the definition.Right. Now my airplane won't drop out of the sky, because the thrust calculation that used nullity as an input produced nullity as an output, in a way completely different from the one that produced NaN from NaN before. This new name and slightly different semantics magically mean the right amount of fuel will go into the engine.
Every calculus student knows the answer to such questions as "what is the limit of 1/x as x approaches 0 from the right" (positive infinity) and "what is the limit of 1/x as x approaches 0 from the left" (negative infinity). And more usefully, derivatives are defined as "the limit of (f(x+h)-f(x))/h as h approaches 0 from the right". Integrals are defined through limits as well. So properly phrased (through limits) questions like this are the very foundation of calculus, and well-understood.
But "1/0" alone? What does that even mean? There's no answer. It's a stupid question, too oversimplified to distinguish between the possible answers. The first thing he wrote - "infinity = 1/0" - was already wrong.
If he made his arguments to his peers instead of schoolchildren, they'd shoot him down, and rightfully so.
Furthermore, saying that computers cannot divide by zero shows a ridiculous lack of common sense. They can do anything we design them to do. Many computer number systems (notably including IEEE 754) have a special value NaN (not a number) that is similar to his nullity concept, except that it's not arrogantly proclaimed as revolutionary or a solution to every problem. Generally, asking a question such as "1/0" indicates a serious logic error. Imagine that airplane needs to calculate the proper elevator trim. Oh, great, the answer is nullity. What does that mean? How should it move the elevators? Giving this failure condition a new name doesn't change the fact that the airplane's still going to drop out of the sky.
Interesting! I'll have to look it over more later.
For comparison, I've put the latest (not yet deployed) version of our offline checkpoint process here. (It's a NetVault backup script; pre locks and does the checkpoint, post touches a file signalling success to our monitoring and releases the lock). It's a procedure outlined by Perforce, though they didn't mention error handling...
Let me be a little more specific: while the hour-long checkpointing process is happening, you can't even open files for edit. In addition to having really course locking, Perforce has more write operations than most version control systems. Subversion's CVS-style working copy means the only write operations are commits and revpropsets.
I use both Subversion and Perforce. There's one major feature still lacking from Subversion: merge tracking. There's work underway to design, implement, and document this feature, but it's not done yet. This is a huge deal for anyone with lots of branches.
Not that it's all roses with Perforce. My impression is that it doesn't scale very well. Most operations simply lock the entire database. I think it's a reader/writer lock, but it means that (for example) while the hour-long checkpointing pre-backup process happens every night, you can't do any write operations. (And there's a way to do an offline checkpoint, but it's not documented or supported, and is difficult to get right, with bad consequences if you don't.)
RedHat's past, present, and future contributions (including to core software such as gcc, the Linux kernel, and glibc) benefit everyone who uses Linux, and nearly everyone who uses any open source software at all. That includes you, even if you are a freeloader, and even if you no longer use RedHat Linux or Fedora.
So they are helping you now. Do they give a fuck about you? Okay, maybe not. I certainly don't; you're a jerk. But that's a useless question - you are taking advantage of their work anyway. Maybe even my work, though you probably don't know it.
There is no such convenant...but if there were, they have not broken it. Formally, the statement "False -> True" is true. You have never scratched their back, and they continue to scratch yours. This makes you a hypocrite, and RedHat a great member of the community. Your "they act like Microsoft" lies do not change that.
You can have an opinion, but I'm done listening to it, and it appears the moderators are as well.
Right. Isn't that what you were doing? Or do you really not see the obvious inconsistency between RedHat's extensive open source contributions and the idea that they don't give a fuck about the community? Is there a third choice? I don't see it...
Close, but that's not the one. I've seen a much more extensive list, with some details about each their involvement in each project, explicitly broken down into the categories I mentioned. I think it was on the Fedora wiki.
No, I foed you immediately after asking for you to be modded down. I don't ever recall hearing of you before; your post just pissed me off.
There never was such a convenant, and RedHat releases all the source packages as required by the GPL. This means you can download a RHEL-based stable, free distribution recompiled by a third party in the form of CentOS. My company tends to buy RHEL where we want paid support or where we have to buy an operating system with hardware (Dell offers Windows or RHEL at the same price). We use CentOS elsewhere. It's the same software, and it's free and legal.
For a company that doesn't give a fuck about open source, they sure hire a lot of people to develop it. I can't find it now, but somewhere there's a webpage with a list of all the open source projects RedHat has developed, maintains, and contributes to. It's amazing how much they've given back to the community, and how many idiots like you there are who are deliberately blind to it.
Can someone please mod the parent down as a troll? And someone else please post a link to the page I can't find?
Reading and re-reading the data should be quicker now, too.
The repository data is stored in a giant XML file which is incredibly slow to parse. Back in the day, it would read this file in every time you ran yum. Last year [1] they added a SQLite cache, so this step could be skipped if the data hadn't changed.
Relatively recently, they added a separate yum-metadata-parser written in C that dramatically reduces the time the parse step takes. Take these changes together and what used to take 45.5 seconds every time you ran yum now takes 7.5 seconds only if the data have changed. [2]
It sounds like they've done as much as they can without changing the transferred data to be an indexed binary format (with the associated forward/backward compatibility complexity).
(I'm not running Fedora Core 6, so I'm not sure if this change made it in.)
[1] - Looks like yum 2.3.1 introduced the cache, around March 2005.
[2] - See this message introducing it around May 2006 sometime after yum 2.6.1.
They couldn't arrest him because he didn't commit a crime. Here's what the article says:
Note the questions it doesn't answer - what business police had checking IDs in this computer lab or forcing people to leave a public building, what law he broke.
You can't arrest someone for a made-up crime...but apparently you can pump over 50 kV through his body five times in rapid succession. Strange world...
Let's assume for a second that I'm a superstar programmer. Why would I want this job? It says nothing! Frankly, sourceforge's website does not appear technically sophisticated, and the existing stuff is all in PHP. What would I be doing? Why would it interest me? Even super-secretive, super-prestigous gives way more information than this. If you want super-star programmers, you're competing with google. You don't have their prestige, so you need to do something better than them. This is it.
You are in good company, though. Much of the time I see places asking for superstar programmers without giving the slightest idea why they're needed or what they'd be doing. It's not a successful strategy...which is probably for the best, because a superstar programmer would be bored out of his (or her) skull if he actually got the job.
I used examples relevant today. If you want older ones, they're essentially unchanged:
Probably less long than you'd think, if any such modem ever existed. 2400 bps modems were 600 baud.
It's khz, small k. Now I understand why you don't see the problem. It causes headaches for people familiar with the SI prefixes outside of computing.
That phrase is nonsense. If they were consistent, they'd be the same, by definition. You previously said that the computer industry was consistent. I showed that only one of three situations within the computer industry used the definition you gave, and you agreed. It's not consistent, and there are many situations where conversions between these three situations causes confusion.
I see that you've gotten a few answers already, so I'll only post mine where they differ. (In other words, my complaints. I'm generally happy with OS X, though - enough so that I ordered a new MacBook Pro when I saw the announcement this morning.)
Pretty good. If your hardware is working, the kernel will be solid.
Finder will be pretty unresponsive if a network drive has stalled - pretty normal for a Unix system, since non-blocking IO is unfortunately not supported on regular files. (The only system I've seen get around this is KDE with its ioslave design.) You can force quit it and it'll come back.
Some stuff takes up annoying amounts of CPU time. Safari will suck down all my CPU (and thus battery life) when on some pages, presumably because of a Flash or JavaScript interaction. Mail + Spotlight really get lost in some massive "Public Folders" hierarchies on my employer's Exchange server. I really wish they had a way of honoring IMAP subscriptions or telling it not to locally cache certain paths. You'll want to run with MenuMeters always visible so you can spot this before discovering that your battery is dead only an hour into the flight.
Fairly easily. The compiler and linker arguments around dynamic libraries are a little weird, but the fink package system (like *BSD's ports) will generally patch things for you, and libtool-based stuff won't have a problem. I basically just type fink install foo and whatever I want shows up.
I don't think their laptops are really any better-constructed than the next guy's. I find all laptops to be pretty fragile - I've had way more problems with them than with desktops. I'm not normally an extended warranty guy, but I'd recommend it on any laptop. AppleCare is quite nice. You can also buy it at any time up to the last day the normal one-year warranty expires.
I didn't know about this feature, but on a later PowerBook, I used the two-finger scrolling. It works wonderfully.
Ignoring for a second that "different" means "inconsistent", you're still wrong - it was never consistent even within the computer industry. Pop quiz:
Use the new definitions. "1 GiB" is much clearer, and eventually people will know that "1 GB" means something else in new software. This sort of switchover is not unprecedented - IIRC astronomy switched to use the same notation for the polarization of light as engineering. Papers around the time would carefully note which definition they used. Ones well before or well after are obvious. More complicated for the computer industry, but not impossible.
[*] - Massive extra credit for correctly considering the protocol overhead for...we'll say FTP over PPPoA VC/Mux over ADSL...but that's not the point of the question.
I see what you're saying, but on the other hand, I'm thinking of a story I read last night. Umm, there's a telling of it here. Short version: Enrico Fermi met some "great" generals and asked what it was that made them great. They decided that it was winning five consecutive battles. And...about 3% of generals were great. Well, that's the same number (.5^5) he would have expected by random chance if all battles were completely even, and all generals were alike.
My point is: maybe the people who gave your interviewer what he was were looking for just got lucky. They might have used the same process (guessing or assuming) as the ones who were wrong. I don't see anything in your story that makes me think that there was enough information to figure out what your interviewer wanted, if only they were smarter. If you ask people to read your mind, some people will get the right answer...but they're still no good at mind-reading.
Though if I were the interviewee, I probably would have done the same thing you did. It seems safer to initially guess real-world rather than contrived. If I say "I wouldn't solve this problem; I'd change it" and they're looking for that, I'm done. If they're not, (1) they'll justify it with a real-world detail and we'll continue, (2) they'll say "this is a contrived interview question; just answer it" and we'll continue, or (3) they'll count it against me, which is just as well because I'll count it against them, too.
Asking questions is good...depending on the question. I've had a few people actually ask "what are you looking for?" The first time, I (stupidly) gave away the whole point of the question, so it became worthless. Other interviewers said that this person was very good at manipulating an interview, but not necessary at the job. Since then, I'll just say "for you to solve this like you would in the real world...except where I specifically tell you otherwise".
Half of that is the interviewer's fault, for not setting expectations. Sometimes you want to see how they'd solve something in the real-world way, and sometimes you want to see how smart/resourceful they are by using contrived situations. "No, you can't ask your coworkers for help; they've all died in a bizarre skiing accident. And you must use this expensive, broken, completely inappropriate product as the basis of your code, because the CTO Has Spoken.[*]" If you made it clear which is which, some people would probably do much better.
I'm starting to wonder if I'm guilty of this myself - my favorite interview question is oversimplified and inherently inefficient. I basically expect them to accept the stupid specification as is (some whining is okay) yet implement it in a real-world way - I emulate "man", "gcc", and "gdb" for them. I try to make that clear, but maybe I could do better...
[*] - Sadly, this second part actually happens. I've seen "you must use Oracle RAC", and I've heard of much worse examples.
I am sure that none of those things happened. Read the KnowledgeBase article I linked to. I've also added to my advogato post the actual partition layout in question.
That seems more in line with the KB article I linked to, but it's a huge difference. This badly violates the principle of "the system should not touch an existing partition of a foreign type without the user's permission". And "reformat unformatted partition?" doesn't count - it should be smart enough to know the answer is no, but at least the question should be "reformat as NTFS extended partition 2 of size XXX that's marked as ext3?"
That's an important observation. Even if you are right about the two prompts (and maybe we have slightly different situations; it sounds like both of your partitions are tagged as NTFS), it's still badly broken. If it's already destroyed your data by the time you know anything is wrong, there's a huge problem.
Oh, by the way:
GNU parted doesn't speak NTFS, but I believe it can do all the Partition Magicy things to a Linux system.