It's not the phone support, it is the errata. And you are mistaken, lots of people still use 7.2/7.3. I have 30 machines to migrate in the next couple of months and our Beowolf cluster has 150+ machines running 7.2.
I've never paid this much. Well, I suppose I have, in that I've bought a day for $10 and only used it for an hour. And yes, I'll pay that (or let my company pay that). If I can get an hours worth of work done in the airport instead of watching CNN, that's a *bargain* for the company. An even better argument is spending that 10 minutes writing an e-mail so someone back at work isn't spinning their wheels for two days.
Often times ARS articles are written at a much higher level than the/. crowd would prefer. Their earlier articles on CPU technology sold me on the site, although I haven't seen much of that kind of reporting in the last year or two. Basically, they *have* been playing down to the slashdot level.
Some (most?) other SLRs have a button you can press that will stop the lens down to the setting it will actually shoot at. This lets you see what will really be in focus and what won't on the final image. It also darkens the image, since less light is coming in, which is why it's not done normally.
It's a very useful feature if you are shooting two different objects at different depths. To use your example, if the final shot was going to be at f/8 and you wanted the background blurry, you don't know if it is going to be blurred or not if you can't preview. Maybe you *have* to go to the f/1.8 or whatever that the lense is at when you are focusing.
Yes, a K-1000 is a good learners camera, but it may not be the best choice for something to grow with. I started with an all manual Olympus camera. Added more lenses, replaced some lenses, added another body, etc. Then I decided I needed AF and something reliable (the Olympus stuff was finicky). I had to scrap everything and buy a new, modern system.
I'd suggest something that can be a seemless upgrade to a modern system. I stuck with my manual equipment far too long because I had too much invested in it.
You want to make sure you pick something you can grow with. From that viewpoint, a
manual focus Nikon body and a couple of lenses might fit your budget. Then later
you can slowly upgrade to auto-focus. A used (or even new) Canon Rebel body and lenses might
also be a good choice. (Nikon kept things compatible when moving from manual to
auto focus, Canon did not.)
Both Canon and Nikon offer digital SLR bodies for when you are ready, and used
equipment is easy to find (unlike some of the other manufacturers).
Keep in mind a few things:
When buying used equipment, cosmetics are very important to the price. So,
if you just want something to use but don't care about resale value, a scuffed
up, but otherwise mechanically perfect, camera or lens may be a great deal.
If you start with an older manual camera, you will learn a lot more about
photography than with a new, auto-everything, camera. You will also waste a lot
more film.
Unless your father had an amazing darkroom, you'll be limited to black and
white prints and maybe developing slides. Color prints are very difficult. But
you'll learn tons regardless.
For what it's worth, I recently replaced an old Olympus system with a Canon
system. Rebel 2000 body, Elan 7e body, 28-90mm lens, and 100-300mm lens. It's
been great. At some point I will buy a digital body too.
Not really. From the crypt manpage there are 7.2e16 possible vales. 2e11 is about 1/300,000 of the possible space. I wouldn't call that substantial, really.
7e16 is of the same order of magnitude as an 8 character password using a combination of letters, numbers, and a few symbols.
Those 207 billion hashes come from only 50 million possible passwords. Using only 10 letters (no upper case) and 8 characters gives 100 million passwords. Bumping the letter pool up to 75 (52 letters, numbers, a few symbols) give you 1E15 possible passwords.
Why should I care what NIC or video card or sound card is in a machine?
We have 30 machines in our research group; there are probably 20 different configurations. Sure, I can find out if I want to, but why should I open each machine up to take inventory before upgrading the OS?
The question is, "Is Fedora going to continue to be RedHat?"
Fedora Core 1 is what was going to be RH10. 90+% of the work was probably already done. Is Fedora Core 2 going to be stable? Is it going to have updates to the important packages? Or is it going to be a dumping ground, a continual "technology preview" that no one can use to get any real work done? How will the effectively nine month lifetime shake out?
All of these are big questions with Fedora and are the reason, where I work, we went to one license of RH Enterprise 3.0 and will either go with Redhat's academic program or a National Laboratory-built version of RH 3.0 which won't cost anything.
I've always found the largest problems with optimizers to be that they sometimes generate incorrect code. He doesn't seem to test for this. Of course, maybe the benchmarks are smaller/less kludgy than the n x 10,000 line codes I've tried to optimize with flags.
Yeah, the fact that it would cost about half a billion dollars and risk the loss of one of the three remaining shuttles is completely irrelevant.
</sarcasm>
I admin a cluster of 30-40 linux boxes in my "spare" time as a researcher.
My point was that the article claimed that nothing like the group of patches that came out for Windows in September had ever, or at least in a long time, happenned to Linux. While the OpenSSH problem wasn't (maybe) exploited, it gave us a flavor of the same thing, having to quickly patch to foil crackers, and patching a piece of software only to find out it was still vulnerable.
BTW, I think I explicitly said it's better than Windows, but that doesn't make it perfect and doesn't mean that Free/Open software doesn't have it its problems that we should acknowledge.
All this article shows is that a thorough analysis is hard, if not impossible. The author goes on with lots of anecdotes and a little searching on CERT. He/she didn't bother to look at Redhat's vulnerability list.
But even doing that, one get's into endless arguments about vulnerabilities in an OS vs. a distribution, severity of vulnerabilities, and whether they are exploitable at the time of the fix.
Also, the author rhetorically asks about a period in Linux as bad as last month for vulnerabilities. I don't admin Windows, but I remember having to update OpenSSH twice in as many days and turn off access to all my users for a few hours because of rumors of an exploit.
So while I think Balmer is full of hot air, let's not paint Linux as a panacea either.
That'd be nice, but it's not crucial for a lot of people. Usually to get a good price on the phone, you sign a one or two year contract. The life of my first phone was a little under two years and when my contract comes up in February, I'm going to want to look at changing phones and providers (and keeping the only phone number anyone knows for me).
Imagine if this thing had hit a towerblock in London, or an apartment complex in New York, possibly killing or injuring hundreds?
Hundreds!?!!? Oh my God!!! It's a good thing we don't have fires anymore, that earthquakes are completely predictable so no one ever dies in those. It's also good we can stop hurricanes off our shores, and 15,000 people don't die in heat waves anymore. So, yeah, now is the time to really get to work on solving a problem that hasn't killed a single person in recorded history.
Slashdot Mirror
It's not the phone support, it is the errata. And you are mistaken, lots of people still use 7.2/7.3. I have 30 machines to migrate in the next couple of months and our Beowolf cluster has 150+ machines running 7.2.
I've never paid this much. Well, I suppose I have, in that I've bought a day for $10 and only used it for an hour. And yes, I'll pay that (or let my company pay that). If I can get an hours worth of work done in the airport instead of watching CNN, that's a *bargain* for the company. An even better argument is spending that 10 minutes writing an e-mail so someone back at work isn't spinning their wheels for two days.
Often times ARS articles are written at a much higher level than the /. crowd would prefer. Their earlier articles on CPU technology sold me on the site, although I haven't seen much of that kind of reporting in the last year or two. Basically, they *have* been playing down to the slashdot level.
How about "Take some of the money you just saved and fund someone to fix it for you?"
It's a very useful feature if you are shooting two different objects at different depths. To use your example, if the final shot was going to be at f/8 and you wanted the background blurry, you don't know if it is going to be blurred or not if you can't preview. Maybe you *have* to go to the f/1.8 or whatever that the lense is at when you are focusing.
Well, a cheap digital SLR with one lens is $1000, for starters. Plus, there is still something to be said for learning with film.
I'd suggest something that can be a seemless upgrade to a modern system. I stuck with my manual equipment far too long because I had too much invested in it.
Both Canon and Nikon offer digital SLR bodies for when you are ready, and used equipment is easy to find (unlike some of the other manufacturers).
Keep in mind a few things:
For what it's worth, I recently replaced an old Olympus system with a Canon system. Rebel 2000 body, Elan 7e body, 28-90mm lens, and 100-300mm lens. It's been great. At some point I will buy a digital body too.
And it's not clear to me that 5e7 (the number of passwords they tried), not 2e11 isn't the right number to compare with 2e11.
7e16 is of the same order of magnitude as an 8 character password using a combination of letters, numbers, and a few symbols.
Moral of the story: Pick a good password.
We have 30 machines in our research group; there are probably 20 different configurations. Sure, I can find out if I want to, but why should I open each machine up to take inventory before upgrading the OS?
Maybe OOo comes with your linux distribution or your XP pre-installed machine and you don't want to fuss with installing Star Office.
That said, I still am using the suite.
Fedora Core 1 is what was going to be RH10. 90+% of the work was probably already done. Is Fedora Core 2 going to be stable? Is it going to have updates to the important packages? Or is it going to be a dumping ground, a continual "technology preview" that no one can use to get any real work done? How will the effectively nine month lifetime shake out?
All of these are big questions with Fedora and are the reason, where I work, we went to one license of RH Enterprise 3.0 and will either go with Redhat's academic program or a National Laboratory-built version of RH 3.0 which won't cost anything.
Ok, you were joking, but there is a serious answer.
Did you not read his post to the mailing list where he says he had words to that effect in the article he submitted but *the editor* took them out?
I've always found the largest problems with optimizers to be that they sometimes generate incorrect code. He doesn't seem to test for this. Of course, maybe the benchmarks are smaller/less kludgy than the n x 10,000 line codes I've tried to optimize with flags.
Yeah, the fact that it would cost about half a billion dollars and risk the loss of one of the three remaining shuttles is completely irrelevant.
</sarcasm>
This is originally from the Onion. Still funny, though.
My point was that the article claimed that nothing like the group of patches that came out for Windows in September had ever, or at least in a long time, happenned to Linux. While the OpenSSH problem wasn't (maybe) exploited, it gave us a flavor of the same thing, having to quickly patch to foil crackers, and patching a piece of software only to find out it was still vulnerable.
BTW, I think I explicitly said it's better than Windows, but that doesn't make it perfect and doesn't mean that Free/Open software doesn't have it its problems that we should acknowledge.
But even doing that, one get's into endless arguments about vulnerabilities in an OS vs. a distribution, severity of vulnerabilities, and whether they are exploitable at the time of the fix.
Also, the author rhetorically asks about a period in Linux as bad as last month for vulnerabilities. I don't admin Windows, but I remember having to update OpenSSH twice in as many days and turn off access to all my users for a few hours because of rumors of an exploit.
So while I think Balmer is full of hot air, let's not paint Linux as a panacea either.
That'd be nice, but it's not crucial for a lot of people. Usually to get a good price on the phone, you sign a one or two year contract. The life of my first phone was a little under two years and when my contract comes up in February, I'm going to want to look at changing phones and providers (and keeping the only phone number anyone knows for me).
Hundreds!?!!? Oh my God!!! It's a good thing we don't have fires anymore, that earthquakes are completely predictable so no one ever dies in those. It's also good we can stop hurricanes off our shores, and 15,000 people don't die in heat waves anymore. So, yeah, now is the time to really get to work on solving a problem that hasn't killed a single person in recorded history.
There is NO way currently to track all the stuff that size in the solar system.