Honestly,just go with the 800 number. I can't say who your Acct Rep would be these days.
Also, I'm not 100% sure that OpenAFS is covered by Sun. I'm pretty sure it is, but... Everything else definitely is covered, though, and at the minimum, you'll get help from a Sun SysEng if you're looking at OpenAFS. (with a Gold Contract, SysEng's respond quite fast.;-)
<disclaimer> I work at Sun. I do not speak for Sun in any way, however. None of what I'm saying is priviledged, or otherwise not publicly available. </disclaimer>
A lot of what is right for you depends on your exact setup. Given your description, I'm assuming you are primarily concerned with file serving for clients, with the possibility of needing to centralize some primary storage for DB or similar app-specific servers. I'm also going to assume you are single-site (given the relatively small amount of storage, I think that's fair).
10TB is not much space (really). I'm going to assume the solution should scale to 3x that over its lifespan (3-5 years). That is, you should have the ability to add up to 20TB or so more rather simply, and without buying major upgrades.
NAS is the cheap way to go, but I'd recommend against it for now: it's rather hard to find one that supports DFS, and OpenAFS isn't well supported either in the NAS space. I'd also shy away from a true SAN solution, since they're going to be way over-engineered for your rather modest needs. By "true" SAN, I'm talking about a large controller head (usually a modified and upgraded FiberChannel switch mated with a small management controller) which front-ends a large number of disk arrays. Rather, I'd recommend a clustered FC solution.
For your problem, I'd look at 2-3 machines which would be your primary file server cluster. They should be hot-clustered together (using your favorite cluster software). I'm going to suggest you use FibreChannel as the back-end direct-attach-storage technology. Connect the head machines to redundant FC switches (you don't need anything really fancy here), and then use JBODs or HW arrays as your storage devices.
Here's a sample solution from Sun (which, if you look at is, is going to be very competative with anyone, including Dell and Build-it-yourself stuff):
(3) Sun x4200 w/ 8GB RAM & 2 dual Opteron 275
(6 total): two single-port 2gbps FC host adapters per server
two 8-port low-end FC switches
(1) 3510FC array w/ redundant HW Raid contollers & 12 x 146GB 10k FC disks
(1) 3511FC JBOD w/ redundant FC connections & 5 x 500GB 7.2k SATA disks
Solaris 10 Update 2 with ZFS (coming Summer 2006)
OpenAFS for Solaris
Samba on Solaris
SunCluster software
Possibly use Zones (i.e. Solaris' VM setup) for better server partitioning
3-year SunSpectrum Gold support (24x7x365 telephone support, 8-8 onsite hardware replacment, 4-hour response time for hardware)
ZFS is the bee's knees. It's just so great. Check it out here: ZFS on OpenSolaris. It's currently available only in the preview Solaris 11 (codename: Nevada), but it will be included in Solaris 10 Update 2 as production-quality code. S10u2 should be available sometime this summer.
I run a similar setup here at work, plus at my private ISP company. The above config is fully supported (all software stacks, hardware, and interaction) by Sun, so you've got 1-stop maintanence support. The above config is for 4.2TB (raw). Assume you RAID-5 the SATA drives, and RAID-10 (striped mirrors) the FC drives, that's 2.9TB usable. Adding additional FC JBODs (either FC or SATA drives) is relatively cheap, and VERY simply to configure.
The solution above is very flexible, and will allow you to add disk and servers to the mix easily and (relatively) cheaply. It is also quite good performance. It does NOT support iSCSI directly to the JBODs or FC switches - you can get an iSCSI HBA for the X4200s should you want to. the 3510FC has built-in administration, so you could attach additional app/DB servers directly to the FC switch for better performance, while still maintaining good overall maintenance/configuration control. The really nice thing about the above hardware config is that it will run Solaris, RedHat or SLES
Honestly, about the only things the new generation of DVD (HD-DVD and Blue-Ray) is going to be a success for is Hi-Def movies. At the size they are, there isn't going to be any demand for them to use on the PC as writable disks, unlike CD-R/W and DVD-R/W. People currently use CDs and old DVDs to do primarily three things: Transfer/backup important data, Audio (whether Orange-book audio or MP3/WMA/AAC), and home-video. All three of these things fit nicely into the current DVD/CD sizes, and even when Camcorders start using HiDef, people generally don't send around multiple hours of Video. At most, it's 1-2 hours of little bobby's Soccer game/birthday party. Which still fits on a DVD via MPEG4 (even in HiDef).
The new DVDs aren't big enough to make an impact on the backup market (where you need 100s of GB per disk to even be considered), and they are (and will remain) far more costly than ordinary CD/DVD-RW media. They have some attractiveness for PC and console gaming, but even there, without a huge amount of in-game video, current DVD capacity will suffice for years for the vast majority of games.
DRM and other factors will hurt uptake even more. Honestly, I figure it's going to take at least 20 years before the new DVD format have anywhere near the penetration that DVDs and VCRs do now. And that takes into account having the new DVD formats on consoles. People just aren't going to use them much.
The big media companies rushed this tech to market - there is no real demand for their functionality right now, and won't be for at least 5 years, minimum. From the consumer standpoint, this is a solution in search of a problem. I figure there will be a generation skip here - the replacement for HD-DVD/Blu-Ray should show up around 2020, and consumers are smart enought to see it, so I'm predicting that the new DVD formats will peak at about 10% of the current DVD market, if that.
I work at Sun, in what used to be called JavaSoft. Just so there's no claim of hidden bias.
The current generation of T1/2000 (called "Niagara") has a major limit, which may or may not impact your app. There is only a SINGLE FPU for each 8-core die. And that FPU doesn't implement the full IEEE floating point specification in hardware.
Every time you need to do an FPU call, it's a major performance hit. So be careful of the app you run on a Niagara-based system. The last word I heard from the Engineering dept was that you wanted to make sure NO MORE than 10% of your instructions were FPU-bound, and that you really should keep the mix to under 5% for best performance. So if you plan on running a home-grown app on it, you might consider source-code changes to eliminate as much floating point math as possible.
All this said, the Niagara makes the conscious decision to be FPU-limited. The market it aims at runs very FPU-light applications - I've seen it run Apache, and boy, does it smoke any x86-based machine when doing so (particularly if your site has heavily database-driven content, in a typical 2-tier setup). Very Fast, Inexpensive, and really Power-Efficient.
Just make sure you ask for a demo machine to test your app with (or talk with a Sun Field Engineer first), so you can find out if Niagara is really for you.
All of this is publicly-available information (go look at the technical documentation Sun has released on the Niagara chip), so I'm not giving anything away that isn't known. Just pointing it out to people to make sure they're happy with the product they buy (we don't want people buying a Pickup and claiming we were supposed to sell them a Ferrari).
That said, the next-generation Niagara (codename "Rock") is already in the works. Talk to you Sun sales rep (or call the 800-number) to find out what's in the works for it, since I can't discuss it here.
You're missing the point. The Fair Housing Act creates "protected" categories, which you (the landlord) cannot use to discriminate.
To quote the Act itself:
Title VIII of the Civil Rights Act of 1968 (Fair Housing Act), as amended, prohibits discrimination in the sale, rental, and financing of dwellings, and in other housing-related transactions, based on race, color, national origin, religion, sex, familial status (including children under the age of 18 living with parents of legal custodians, pregnant women, and people securing custody of children under the age of 18), and handicap (disability).
Thus, it protects people from being discriminated on basic, common things which we generally accept are BAD for society to allow others to penalize them for.
Now, it says nothing of Political viewpoints or many other behavior attributes, nor does it prohibit exclusion on profession; it also has allowances for large dwellings which provide a common experience for a narrower-than-normally-allowed populance: thus, single-sex and single-religion housing developments are allowed, but generally under very limited and restricted conditions, and certainly not for anything that doesn't have a moderately large number of units. Thus, if you were the owner of 10 apartment complexes, each of 100 units, it would probably be permissible to designate one of the them for Catholic Men only (for instance). However, your milage will vary (and so will local ordinances).
Without the FHA, it becomes easy to lock various ethnic/religious groups out of housing, using nothing more than social convention. And that is BAD. See the pre-Civil-Rights era for any example you want.
Free speech cannot trump everything - this is one of those cases where allowing everyone to do what they want oppresses a minority, whether they (landlords) specifically intend to or not. It is similar to requiring public businesses to serve anyone who walks in - McDonald's can't refuse to serve blacks, no matter what the owner of that McDonald's would like.
I'm not sure how the ruling will classify (no pun intended) the ads on Craigslist - though I suspect that since there is no fee involved, Craigslist will have Common Carrier or equivalent status - I'm all for holding on-line sites to the same standard as print newspapers, so long as it TRULY the same standard.
That said, the relevant sections of the Fair Housing Act do NOT apply to ads for roommates or those looking to share a place they do not own. The law involves the Owner (or the duly-empowered representative) and any prospective lessee. I'm a little fuzzy on sub-leasing, but since most of that is a huge gray market anyway (most rental contracts forbid subleasing, but it's commonly ignored), I suspect that it isn't covered in the F.H.A. Note that if you own the place, renting out the spare room does make you the landlord, and you have to abide by the F.H.A. But if you're a renter, and just looking for a roommate, well, the F.H.A. doesn't apply, and you can specify that only Left Handed, Purple Skined Demon Sycophants can apply to be your roommate.
Most likely, the suit will get a summary judgement and be punted. But it at least should make those services which DO charge think about complying with the F.H.A. Which is only, well, fair.
As it has been said before, it sounds like the OP did the ethical and professional thing in his resignation, and the company opted for the (now fairly standard) rude and unprofessional immediate termination. That said, everyone should know what your state's employment laws are. They vary widely, and give the employee a variety of options and rights, and also can help you set expectations.
I'm going to speak about California, since that's where I work now. IANAL, but I've talked to one about this, and you should too. It's cheap ($100 or so for 30 minutes or so), and will give you information that is very much worthwhile, both at the start phase (negotiating your employment) and exit phase (termination/resignation) of your job.
CA is an "at will" state. For those employees (not contractors) not covered by a union contract, there are really three different ways to end employment:
Termination for Cause - your employer decides to fire you, and cites one of a limited number of state-specified reasons for doing so. Generally, "for cause" is limited to (documented) bad behavior on the employee's part. Usually not criminal behavior (criminal behavior at the company falls within "for cause", however), but for things like repeated violations of confidentiality, perpetual tardiness, etc. This is quite narrow, and the employee generally has to have a documented trail of bad behavior, and been formally warned about it by the company. Termination for Cause can be done at any time, is effective immediately with no notice, and the ex-employee DOES NOT have the right to State Unemployment Benefits.
Termination without Cause - the company decides they don't want you for a reason other than one that falls under "for cause". It can be simply that your job isn't needed anymore, you pissed off the CEO, you don't seem to have the skills for the job, they don't like the color of your shirt, etc. Within the first 3 months of your employment with the company, they can fire you at any time, with no notice, and your employment ends when they notify you. After 3 months of employment, 2 weeks notice of termination of employment is required. In either case, you qualify for State Unemployment Benefits after leaving.
Resignation - the employee decides to quit. This can be done at any time, for any reason. The employee is REQUIRED to give 2 weeks notice as to the date they will cease being an employee. Failure to do so can be considered "work abandonment", and can be reasons to be fired under "for cause". Of course, since you're quitting anyway, it's seldom an issue. Employees quitting are not eligible for State Unemployment Benefits.
Now, what happens often these days is that the company notifies you that you are terminated, and then tells you (e.g. locks you out, etc) that you are not to come to work for the next 2 weeks. The same applies to people resigning when they give notice (as the OP found out). HOWEVER, you are STILL CONSIDERED EMPLOYED by the company until the 2 week period is up. This is often important for Stock vesting, etc. And don't let them fool you that the "2 weeks pay" thing is a "severance package". It isn't. They are REQUIRED to pay you as long as you are an employee.
Don't Ever Sign Any End-Of-Employment Contract To Get a 2-Week Severance. You're an idiot if you do - they owe you the money in any case. The only time you should sign one of the agreements is to get money beyond what would be coming to you AFTER YOUR RESIGNATION/TERMINATION DATE (not the date you gave/received notice).
As a side note, this idiotic "walk-them-out-when-they-resign" policy seems to have originated in Silicon Valley in the 90s, as a consequence of the Dot-Com boom. Too many companies with no proper HR department not having any sort of a clue as to how to professionally hire/fire people. Unfortunately, it seems to have become a trend (it's the norm here in Silicon Valley for everyone, including the huge companies), which is telling as to the lowered quality of management (and HR) of companies these days.
One of the biggest productivity saps for a sysadmin is dealing with the massive volume of email that we get. In even a moderate-sized business, it's easy to get 1,000 mails per day, with a couple 100 actually from a person, not an automated script.
Now, what I'm about to say is predicated on the assumption that your external mail server already runs SPAM filters, and that virtaully everything that you actually get is "real" mail. If this is not the case, FIX THIS FIRST. Get your company to pony up for some serious anti-spam software. It saves EVERYONE a ton of time, and at the same time, cuts down on your (the company's) exposure to the nasties that inhabit email.
First, pick an email client which has filters. My preference is for Evolution or Thunderbird, but there are many out there. Pick one. As a previous poster noted, GIVE IT ITS OWN DESKTOP - that is, in your window manager which has virtual desktops, dedicate one solely for the email client. Now, configure it with lots of filters to sort your mail. Personally, I have a reasonable hierarchy with 3 folders at the top level: NOW, LATER, and WHENEVER. Underneath these, there should be folders for every type of email you get: ones from your boss, ones from the company HR, ones from the monitoring scripts running on your servers (you do have these, right? RIGHT?). Take a good long time figuring out how to get these down cold - you want a good balance of sufficient sorting without going overboard. I find that having about 30-50 folders total is optimal for me. If you can, also have the email client tag your mail with "importance" color coding (most clients have this, and it's really useful).
Now to reading: obviously, your should read the NOW, well, NOW. However, you don't want to be completely interrupt-driven. I would turn off any biff-style mail notification, or at best, turn down its check time to no less than 10 minutes between check. Instead, train yourself to periodically check the NOW folder. Read and deal with the NOW stuff during your normal workflow.
The LATER folder should probably be read every couple of hours, or if you truly haven't anything else to do. Resist the temptation to open it and look. Finally, the WHENEVER shouldn't be read until the end of the day (or maybe while your eating lunch at your desk;-)
Email is one of the great things about networks; however, it can be an enormous timesink if not properly handled.
-Erik
I've noticed everyone here is recommending their favorite language, without mentioning why it might be useful for beginning programming. I'd like to address that issue:
In Beginning Programming, you're assuming that the student know absolutely nothing about programming. They should know Algebra (to understand variables), but that's about the extent of Math.
Thus, probably the Right Thing to do is build on the basic Math understanding that your students have, and start to introduce programming concepts from there.
For that reason, I'd start with a Functional Programming language as they tend to be rather obviously Math-derived, and help ease the introduction of programming constructs. I would recommend Scheme, since it is associated with one of the best "Teaching Programming" texts ever, Structure and Interpretation of Computer Programs.
After they've grasped Functional Programming, I'd have them move on to Procedural Programming, starting in Pascal or Modula-2. These two are excellent languages which illustrate the fundamentals of procedural programming, without some of the nastier pitfalls. There are also excellent textbooks available for Pascal (fewer for Modula-2). Later, I'd move them to C to introduce pointers and some of the other hairier features.
Finally, they're ready for Object Oriented Programming, for which I'd use Java - it's widely used, very common for college-level coursework, and there are a large number of supporting utilties and good textbooks useful for teaching aids.
Overall, I'd look at teaching Functional in 1 semester, Procedural in 1 or 2 semesters (depending on the detail and breadth you want), and OO in 1 or 2 semesters.
In all honestly, I love scripting languages like Perl, Python, and Ruby. However, as a teaching tool, they're all too multi-purpose, and it's easy for the student to do something they're not supposed to do (even though it works). For teaching languages, you want ones which pretty much only allow the student to program in the methodology you're teaching. That is, you generally want those languages which are LESS flexible, since your main goal is correctness, not functionality.
-Erik
Multicore is great, but not for the obvious reason
on
AMD Quad Cores, Oh My
·
· Score: 4, Informative
Yes, Virginia, we can use mutli-core. I mean, we're all into SMP heavily in the non-desktop role (does anyone actually make a "server" that doesn't have SMP?)
There are two big things I love about the multi-core Opterons: They draw less power than equivalent SMP machines (acutally, quite a bit less), and they allow multiple "CPUs" to use the same memory controller. Nominally, the second isn't a big win, but it can be for practical purposes.
Opterons have dedicated memory channels on them, so a current dual-socket Opteron has two DISTINCT DIMM banks - that is, on a motherboard with 8 DIMM sockets, 4 are allocated to each CPU socket. So if you have only one CPU, you can only use 4 DIMM sockets. Since those 4 sockets are often configured as a single bank (i.e. they all have to be filled to work), you can't add another CPU to the system without buying more RAM. This is wasteful. But with a multi-core opteron, all on-chip cores share the same memory bank.
The jist of this is that it'll be easier to have High-Compute, lower RAM configurations than it currently is reasonable to do. There are a lot of tasks out there which it is really nice to have a modest amount of RAM (say 4GB), but with huge crunch. Currently, it's hard to buy a config to do that, since you generally either end up way over-paying for CPUs, a huge number of tiny DIMM chips (which sucks for future expansion), or a larger number of motherboards, which draws more power.
And, hey, they're not tooo bad in price. Sun's dual-core v40z is less than twice as expensive as their single-core v40z, and you save lots on power/cooling/space.
I see a lot of people decrying the 9-year length of sentence as excessive. I'd like to promote the idea that its actually lenient, given the harm to society.
First, for those who haven't RTFA, this guy's crime wasn't just "spamming", it was the electronic equivalent of mail fraud. Take a look here for mail fraud penalties. Yup - that's right. Up to 5 YEARSper occurance. Not per person actually defrauded, but per mail sent.
Furthermore, we seem to want to punish "blue collar" crime (physical violence and theft) as somehow more heinous than "white collar" crime (usually fraud and theft of money or intangible property). As a poster above noted, blue collar crime tends to have a severe impact on a very limited number of people, though in the aggregate it also attacks a locality's social fabric (consider the number of violent crimes in someplace like South Central LA and the correlation to property values there). White collar crime, however, tends to impact a large number of people to a lesser extent, but also directly attacks the fundamental underpinnings of the society: in particular, the fundamental trust in fairness and shared responsibility that is essential for modern societies to function.
Fraud in particular is a particularly heinous crime from a societal standpoint, as it attacks the basic trust we put in financial transactions. A CEO giggering quarterly numbers is doing more than just cheating some stockholders out of a few cents in stock price - he's attacking the whole investing system which depends on truth in information dissemination. For if investors can't trust that a company's 10k annual report has real numbers, how can they invest?
White collar crime needs to be far more heavily punished than it currently is. And, it is much more deterred by increased prosecution and higher penalties than blue collar crime. Blue collar crime is generally only deterred by increased police presence (i.e. preventative measures) and not by increased penalties. White collar crime, on the other hand, generally shows a strong correlation to the likelihood of prosecution and severity of penalty. This is due to the fact that most white collar crime is committed by the more wealthy segment of the population, who generally do a risk analysis before committing the crime (i.e. "I'll steal $100,000 from the company, if I'm only 10% likely to get caught and only face 3 months in jail, but won't steal if I've a 50% chance of getting caught or if the sentence will be 5 years").
Also, remember that as "non-violent" criminals, white collar criminals tend to get put in low-security prisons, which cost much less to maintain than those in for violent crimes.
Overall, I'd like to see us start to put the emphasis on white collar crime instead of blue collar crime. In the big picture, I think it's far more damaging to society, and is far more frequent than people think.
There are some issues with this case (more specificly, the technicalities of the anti-spam law), but in the big picture, I think the sentence is exceedingly fair.
As cool as the DMC-based Time Machine is (and I have to admit, the original B-T-T-F movie is a good memory of my teenage years), the whole BB stuff just rocked. Soooo much more wacked, and so much more fun.
Honestly, if I were a games developer looking at the Linux market seriously, there is one feature which would really draw me in: the ability to provide a bootable distribution on the game CD.
One of the biggest headaches of game developers is trying to test their game on a sufficiently large subset of available hardware and software configurations to insure it will work properly. This isn't an issue on Consoles, which is one (not the only, but a big one) of the issues they are so popular to develop for. Having a bootable distro on the game CD gives the developer many of the advantages of both Console and PC:
Known Software Config - by using a bootable CD, the developer can pick exactly the software versions wanted, and not have to settle on some generic baseline likely to be available to the whole audience
Simple QA - with a known software config, QA is vastly simplified. The software will only support specific hardware, so any other hardware configs don't have to be tested. And multiple versions of the same software are no longer an issue. All of this results in much higher quality product.
Moore's Law - unlike Consoles, the PC market allows for quick hardware turnover. So a fixed CD distro can freeze the software config, while allowing more advanced hardware to improve gameplay.
Permanent Storage - Given a properly written CD distro, saving config data and game state to permanent media is rather simple (autodetect the hard drive, and then ask the user to tell where to save it). This is a big win over the Console, where saving state is restricted due to cost and available space.
Given the size of modern games, DVD distros are more likely than CD distros, but the concept is identical.
The bootable game CD/DVD has the potential to drastically reduce developer costs associated with modern games, and merge the best features of PC and Console gaming, with few drawbacks. I expect to see game makers venture into Linux in this area first.
As an aside, 1 cubic meter of helium provides about 1 kg of lift (bouyancy). Hydrogen is only slightly better at about 1.2 kg per cubic meter.
So, if you want to hover a modest size craft (say 1 metric ton or so, which is rather small), you need to use about 1000 cubic meters of helium. Which is a bag 10 meters per size. Not a chance in hell for his design.
For example: a modern 737-800 aircraft weighs roughly 40 MT empty, has a cabin size of 40m long by 3.5m in diameter, and has a maximum cargo weight of about 30MT. Given that you can reduce the craft weight by 75% with advanced materials (very optomistically), you still need a gas volume of 40,000 cubic meters to lift the craft and cargo. This occupies a space 104 times the main cabin volume.
Yes, he talks about a balloon-like structure. More specifically, a Zeppelin-like structure, with a rigid airframe holding the lighter-than-air gases. However, the design is of a plane-like glider. Take a look at a modern zeppelin. The radio of volume to surface area is many times greater than what he's proposing. Essentially, he's using lifting gases to enhance the performance of the craft. No matter what he's building the craft out of, it won't be able to hover on lift-gas alone, given his design.
The bottom line here is this: either you have to build a balloon/zeppelin-like craft with a large enough gas resevoir (in which case the aerodynamics of the structure completely preclude gliding), or you build a glider which uses wing lift to stay aloft. It's either or - you can't have both. The only way I could possibly think to provide enough lift for hovering in a glider is to have a hard vacuum instead of a helium (or hydrogen). But that would require a considerably stronger container, which in turn raises the total weight, Q.E.D.
He's designed an aeroplane, despite his attempts at misdirection.
Honestly, after reading this site, it smacks of all the hype around the Segway (and many similar, less-successful "revolutionary breakthrough" schemes).
Yes, it is possible to create a fuel-less aeroplane that can still maintain forward motion. Advanced glider technology certainly fits this description. However, there are a couple of things that are missing from the adware:
How do you fly into the wind? Without some serious motive force, traveling against even moderate air-currents is impossible, or can be done at such a slow rate as to render travel unusable.
Gliding requires a very high lift-to-weight ratio. That is, you generally need a very large wing area to lift even a small amount of weight. And, of course, the wing weight contributes to the overall weight of the craft. The result is a very, very, very small cargo capability.
The efficiency of any wind-generator (even a revolutionary one) would never outstrip detrimental effects of drag it produces nor the loss of cargo space that the weight of the generator occupies.
Winged aircraft cannot hover without some form of downward thrust. Basic aerodynamic physics here. Winged aircraft depend on forward motion to provide lift, and thus the ability to fly. The ability to hover requires one of three things: (1) the entire craft has positive bouyancy (i.e. blimp/balloons), (2) a moving wing providing its own lift (i.e. helicoper rotors), or (3) downward air thrust (i.e. Harrier and similar). The craft described has none of these.
The ability to climb in an aerostatic craft requires favorable air currents, and a minimum forward velocity. The first condition is highly unpredictable, and generally not present for hours or days, depending on location. The second condition has to be provided by either motive force (e.g. engines) or gravity (which is why gliders are often launched from cliffs).
The physics of underwater motion are similar those of flight - basic fluid dynamics here. The problem is he's ignoring fundamental environmental differences between water and air. The density differences between the two make it possible to move large masses underwater, but only tiny masses in the air using the same principles. Not to mention that the fluid consistency and motion between air and water are radically different, which invalidates using the ocean as a model for the sky.
Now, I usually don't get this annoyed, but we've gone through a bunch of these recently, and I'm sick of it.
The GPL is not Holy Scripture
The GPL is nice. It works for a large amount of stuff. However, it very much does NOT work for many other things, even in the Open Source world. I don't (and you shouldn't) want everything to be GPL'd. The GPL has a LOT of problems, freedom to copy aside. It is definitely not suitable for every purpose, given our current legal framework.
We should measure a license by how obnoxious and restrictive it is, not some idiot litmus test of GPL-compatibility. I prefer that we gravitate to a small number of general licenses for simplicity's sake, but there is no real good reason that they all HAVE to be GPL-compatible, any more than they all HAVE to be BSD-compatible.
The various ASL versions are all very benign and nice BSD-ish licenses, that may or may not be GPL compatible. They have very liberal code reuse and copying provisions, and very few restrictions. If they are GPL-incompatible, well, then, that's life. I'm not going to get angry over this, any more than I get upset because I can't use GPL libraries with my proprietary code.
Please quite focusing on the idiotic minutia, and pay attention to the hard issues of license lock-in and IP coralling prevelant in software licensing today.
So far, the FCC and Congress haven't done much to regulate the VoIP and other technologies which will make VideoPhones ubiquitous in the coming years. (yeah, it's coming, whether we're all ready or not...)
I'd really rather NOT have them regulate this new area in the same manner which they regulate traditional telephony. However, I'm not in favor of the complete hands-off approach either. I'm hoping to see some sanity here - we certainly need a Universal Access Fee structure for Internet connections - if we want internet connections to be as common as telephones (and I think we can all agree that this is a Good Thing for the country), then we need to do a bit of subsizing.
Think of it this way: we plan to use Internet connectivity to Vote, use VideoPhone, standard Internet Telephony, and a host other things which we pretty much are going to designate as ESSENTIAL to living in the U.S. in a decade or two. Thus, we're going to have to subsize certain minimum connectivity offerings for those not able to afford the whole thing. Look in the front of your Telephone Book to see what I mean for current telephone service.
If we're lucky and a bit vigilant over the legislation being proposed, we can get the Right Thing done: basic Universal Service for everyone, while replacing the horrid mis-mash of telephony regulations with something much simpler and easier to deal with. If we're not, well, guess what - VideoPhone is going to be regulated even worse the regular telephony.
I'd agree with the poster above that this was a "creative" use of the Carrier law by an FBI agent hoping to bully the guy into complying, with little actual legal ground to stand on. However, the reporter is definitely going to get a request for the full notes and info from the interview.
As the Times article points out, a reporter has no LEGAL means of keeping these away from the FBI, if really pressed to present them. We'll see if it gets that far.
Section 4.3.1 of RFC 1034 pretty clearly states that the response to a name query is to be:
If recursive service is requested and available, the recursive response to a query will be one of the following:
The answer to the query, possibly preface by one or more CNAME RRs that specify aliases encountered on the way to an answer.
A name error indicating that the name does not exist. This may include CNAME RRs that indicate that the original query name was an alias for a name which does not exist.
A temporary error indication.
If recursive service is not requested or is not available, the non-recursive response will be one of the following:
An authoritative name error indicating that the name does not exist.
A temporary error indication.
Some combination of:
RRs that answer the question, together with an indication whether the data comes from a zone or is cached.
A referral to name servers which have zones which are closer ancestors to the name than the server sending the reply.
RRs that the name server thinks will prove useful to the requester.
Now, the section thereafter goes on to talk about wildcards, so they are pretty much out of luck for saying that VeriSign isn't implementing the RFCs correctly. However, another portion of the RFC makes it very clear that wildcards are only for use within an entity's domain of control (that is, *.foo.com in DNS will not affect lookups under bar.com). The key here is that it is up to the OWNER of the domain in question as to the appropriateness of wildcards in DNS. VeriSign does NOT OWN THE.COM TLD. They merely ADMINISTER it for ICANN. Thus, there is a very good case for VeriSign being in breach of contract by failing to cary out the wishes of the OWNER of the.COM TLD. Which in this case is ICANN.
Basically, I would be a bit more thorough before going to VeriSign, but afterwards, I'd still wack them over the head with the contract and force them to remove the wildcard.
I see a couple of people above moaning that SBC is forced to sell the "last-mile" loops to competitors for under cost, meaning the price they charge to the CLEC is less than it costs to maintain the wire.
This is patently false.
What the truth of the matter is this: the ILEC has to price the loop costs equally for all comers, INCLUDING THEIR OWN INTERNAL CUSTOMERS. Thus, out here in PacBell land, the costs that PacBell charges a CLEC (say AT&T or Covad) to lease a loop is the same costs is must charge PacBell Internet to lease that loop.
Guess, what? The ILEC like to subsidize their ISP and premium service groups by "selling" them loops for less than they cost. Regluation simply forces the ILEC to play fair, by allowing other CLECs to get this same price, and thus not allowing the monopoly on physical loop ownership to spill over into other services.
The ILEC could charge CLECs the proper amount to cover their costs, but they'd have to charge their in-house divisions the same rate. Thus, in reality, it is not the CLECs who are getting the free ride on the backs of the ILEC, but that the ILEC is propping up one of its own companies at the expense of another part of the ILEC conglomerate.
The Kansas legislature was completely correct - don't ever believe an ILEC "promise" in exchange for relaxation of regulation. They lie through their teeth constantly.
Fundamentally, the real solution is to force the ILECs to divest from physical loop ownership, and spin off a seperate company which is only allowed to own the loops, but may not sell data/voice services over those loops. Keep the hardware (a natural monopoly) distinct from the data (a natural competive market). Right now, we mix the two, to the detriment of all.
First off, there is no need for the captain to be "instantly reachable". It's not like he's the only one which can make command decisions on a ship. It's been a while, but IIRC the title of the person who is in control of the ship is the Officer of the Deck. Should neither the captain nor Executive officer be on the bridge, one officer is designated the OOD and has effective command of the ship. Now, in a crisis, the XO and Captain almost always attempt to return to the bridge to reassert command, but the OOD can make all decisions (including breaking previous captain's orders, should the OOD deem it necessary) until relieved. So, it is silly to design a system to allow the captain to controll the ship from anywhere. Someone in the chain of command is already doing that from the place most suitable to do so, the bridge (or CIC, as appropriate).
Second, virtually all ships have a voice intercom systems set up throughout, which can relay orders back to the bridge far faster and more efficiently than some silly handheld WAP thingy. They're hardwired, so no emissions. They are invariably redundant, and far more likely to survive damage than a WAP system.
Finally, reduced manpower is a great goal, but generally is highly driven by putting in machinery which requires fewer operators. Communications systems are not really any manpower saver. And, as noted by others, you need twice as many people on a ship as it takes to operate all machinery: remember you have to run the ship 24x7, so you need at least two shifts (there's a little overlap, but 2x is a good rule of thumb), and you better have some extras for damage control and casualty replacement. So, you'll get manpower savings by automatic ammunition loading systems, better fire-supression, more efficient engines, better EW weapon systems, but not by adding WAP points.
While they're technically a pure commertial venture, and thus I'm not exactly sure how closely they mimic the ClearChannel-style monopoly of "you-will-listen-to-this" music, I find that the BBC radio channels are an excellent alternative source of music.
They have several HUNDRED streaming media files available for a huge variety of music, much of which is hard to find in the USA. I find it very refreshing, since British musical taste as always been a bit different than American taste, though the two tend to be quite compatible.
All their DJ shows are available on-line, and are updated weekly. And, since it's the BBC, they have virtually anything that you might be interested in, from Talk Shows about Scottish Sheepherding to the latest Dance club hits from the Continent.
Actually, according to the FCC here in the USA, "Broadband" has a very specific meaning. That is, the download speed must be at least 208kbps. This is a legal definition. You can't sell a ISDN line and call it Broadband.
In general, that's a good definition from a networking standpoint, so I tend to stick with it in any case.
Oh, and the origin of the term "broadband" had NOTHING to do with the ability to send multiple signals on a single wire. It has to do with the SPECTRUM used to send those signals. Standard voice-only telephony uses a relatively small frequency range (about 4kHz). Broadband was originally used to describe an analog signal that used several multiples of this spectrum to send the data. Thus, it was much "broader" than the voice spectrum.
Slashdot Mirror
Honestly,just go with the 800 number. I can't say who your Acct Rep would be these days.
Also, I'm not 100% sure that OpenAFS is covered by Sun. I'm pretty sure it is, but... Everything else definitely is covered, though, and at the minimum, you'll get help from a Sun SysEng if you're looking at OpenAFS. (with a Gold Contract, SysEng's respond quite fast. ;-)
-Erik
OK. Up front:
<disclaimer>
I work at Sun. I do not speak for Sun in any way, however. None of what I'm saying is priviledged, or otherwise not publicly available.
</disclaimer>
A lot of what is right for you depends on your exact setup. Given your description, I'm assuming you are primarily concerned with file serving for clients, with the possibility of needing to centralize some primary storage for DB or similar app-specific servers. I'm also going to assume you are single-site (given the relatively small amount of storage, I think that's fair).
10TB is not much space (really). I'm going to assume the solution should scale to 3x that over its lifespan (3-5 years). That is, you should have the ability to add up to 20TB or so more rather simply, and without buying major upgrades.
NAS is the cheap way to go, but I'd recommend against it for now: it's rather hard to find one that supports DFS, and OpenAFS isn't well supported either in the NAS space. I'd also shy away from a true SAN solution, since they're going to be way over-engineered for your rather modest needs. By "true" SAN, I'm talking about a large controller head (usually a modified and upgraded FiberChannel switch mated with a small management controller) which front-ends a large number of disk arrays. Rather, I'd recommend a clustered FC solution.
For your problem, I'd look at 2-3 machines which would be your primary file server cluster. They should be hot-clustered together (using your favorite cluster software). I'm going to suggest you use FibreChannel as the back-end direct-attach-storage technology. Connect the head machines to redundant FC switches (you don't need anything really fancy here), and then use JBODs or HW arrays as your storage devices.
Here's a sample solution from Sun (which, if you look at is, is going to be very competative with anyone, including Dell and Build-it-yourself stuff):
ZFS is the bee's knees. It's just so great. Check it out here: ZFS on OpenSolaris. It's currently available only in the preview Solaris 11 (codename: Nevada), but it will be included in Solaris 10 Update 2 as production-quality code. S10u2 should be available sometime this summer.
I run a similar setup here at work, plus at my private ISP company. The above config is fully supported (all software stacks, hardware, and interaction) by Sun, so you've got 1-stop maintanence support. The above config is for 4.2TB (raw). Assume you RAID-5 the SATA drives, and RAID-10 (striped mirrors) the FC drives, that's 2.9TB usable. Adding additional FC JBODs (either FC or SATA drives) is relatively cheap, and VERY simply to configure.
The solution above is very flexible, and will allow you to add disk and servers to the mix easily and (relatively) cheaply. It is also quite good performance. It does NOT support iSCSI directly to the JBODs or FC switches - you can get an iSCSI HBA for the X4200s should you want to. the 3510FC has built-in administration, so you could attach additional app/DB servers directly to the FC switch for better performance, while still maintaining good overall maintenance/configuration control. The really nice thing about the above hardware config is that it will run Solaris, RedHat or SLES
Honestly, about the only things the new generation of DVD (HD-DVD and Blue-Ray) is going to be a success for is Hi-Def movies. At the size they are, there isn't going to be any demand for them to use on the PC as writable disks, unlike CD-R/W and DVD-R/W. People currently use CDs and old DVDs to do primarily three things: Transfer/backup important data, Audio (whether Orange-book audio or MP3/WMA/AAC), and home-video. All three of these things fit nicely into the current DVD/CD sizes, and even when Camcorders start using HiDef, people generally don't send around multiple hours of Video. At most, it's 1-2 hours of little bobby's Soccer game/birthday party. Which still fits on a DVD via MPEG4 (even in HiDef).
The new DVDs aren't big enough to make an impact on the backup market (where you need 100s of GB per disk to even be considered), and they are (and will remain) far more costly than ordinary CD/DVD-RW media. They have some attractiveness for PC and console gaming, but even there, without a huge amount of in-game video, current DVD capacity will suffice for years for the vast majority of games.
DRM and other factors will hurt uptake even more. Honestly, I figure it's going to take at least 20 years before the new DVD format have anywhere near the penetration that DVDs and VCRs do now. And that takes into account having the new DVD formats on consoles. People just aren't going to use them much.
The big media companies rushed this tech to market - there is no real demand for their functionality right now, and won't be for at least 5 years, minimum. From the consumer standpoint, this is a solution in search of a problem. I figure there will be a generation skip here - the replacement for HD-DVD/Blu-Ray should show up around 2020, and consumers are smart enought to see it, so I'm predicting that the new DVD formats will peak at about 10% of the current DVD market, if that.
-Erik
I work at Sun, in what used to be called JavaSoft. Just so there's no claim of hidden bias.
The current generation of T1/2000 (called "Niagara") has a major limit, which may or may not impact your app. There is only a SINGLE FPU for each 8-core die. And that FPU doesn't implement the full IEEE floating point specification in hardware.
Every time you need to do an FPU call, it's a major performance hit. So be careful of the app you run on a Niagara-based system. The last word I heard from the Engineering dept was that you wanted to make sure NO MORE than 10% of your instructions were FPU-bound, and that you really should keep the mix to under 5% for best performance. So if you plan on running a home-grown app on it, you might consider source-code changes to eliminate as much floating point math as possible.
All this said, the Niagara makes the conscious decision to be FPU-limited. The market it aims at runs very FPU-light applications - I've seen it run Apache, and boy, does it smoke any x86-based machine when doing so (particularly if your site has heavily database-driven content, in a typical 2-tier setup). Very Fast, Inexpensive, and really Power-Efficient.
Just make sure you ask for a demo machine to test your app with (or talk with a Sun Field Engineer first), so you can find out if Niagara is really for you.
All of this is publicly-available information (go look at the technical documentation Sun has released on the Niagara chip), so I'm not giving anything away that isn't known. Just pointing it out to people to make sure they're happy with the product they buy (we don't want people buying a Pickup and claiming we were supposed to sell them a Ferrari).
That said, the next-generation Niagara (codename "Rock") is already in the works. Talk to you Sun sales rep (or call the 800-number) to find out what's in the works for it, since I can't discuss it here.
-Erik
You're missing the point. The Fair Housing Act creates "protected" categories, which you (the landlord) cannot use to discriminate.
To quote the Act itself:
Thus, it protects people from being discriminated on basic, common things which we generally accept are BAD for society to allow others to penalize them for.
Now, it says nothing of Political viewpoints or many other behavior attributes, nor does it prohibit exclusion on profession; it also has allowances for large dwellings which provide a common experience for a narrower-than-normally-allowed populance: thus, single-sex and single-religion housing developments are allowed, but generally under very limited and restricted conditions, and certainly not for anything that doesn't have a moderately large number of units. Thus, if you were the owner of 10 apartment complexes, each of 100 units, it would probably be permissible to designate one of the them for Catholic Men only (for instance). However, your milage will vary (and so will local ordinances).
Without the FHA, it becomes easy to lock various ethnic/religious groups out of housing, using nothing more than social convention. And that is BAD. See the pre-Civil-Rights era for any example you want.
Free speech cannot trump everything - this is one of those cases where allowing everyone to do what they want oppresses a minority, whether they (landlords) specifically intend to or not. It is similar to requiring public businesses to serve anyone who walks in - McDonald's can't refuse to serve blacks, no matter what the owner of that McDonald's would like.
-Erik
I'm not sure how the ruling will classify (no pun intended) the ads on Craigslist - though I suspect that since there is no fee involved, Craigslist will have Common Carrier or equivalent status - I'm all for holding on-line sites to the same standard as print newspapers, so long as it TRULY the same standard.
That said, the relevant sections of the Fair Housing Act do NOT apply to ads for roommates or those looking to share a place they do not own. The law involves the Owner (or the duly-empowered representative) and any prospective lessee. I'm a little fuzzy on sub-leasing, but since most of that is a huge gray market anyway (most rental contracts forbid subleasing, but it's commonly ignored), I suspect that it isn't covered in the F.H.A. Note that if you own the place, renting out the spare room does make you the landlord, and you have to abide by the F.H.A. But if you're a renter, and just looking for a roommate, well, the F.H.A. doesn't apply, and you can specify that only Left Handed, Purple Skined Demon Sycophants can apply to be your roommate.
Most likely, the suit will get a summary judgement and be punted. But it at least should make those services which DO charge think about complying with the F.H.A. Which is only, well, fair.
-Erik
As it has been said before, it sounds like the OP did the ethical and professional thing in his resignation, and the company opted for the (now fairly standard) rude and unprofessional immediate termination. That said, everyone should know what your state's employment laws are. They vary widely, and give the employee a variety of options and rights, and also can help you set expectations.
I'm going to speak about California, since that's where I work now. IANAL, but I've talked to one about this, and you should too. It's cheap ($100 or so for 30 minutes or so), and will give you information that is very much worthwhile, both at the start phase (negotiating your employment) and exit phase (termination/resignation) of your job.
CA is an "at will" state. For those employees (not contractors) not covered by a union contract, there are really three different ways to end employment:
Now, what happens often these days is that the company notifies you that you are terminated, and then tells you (e.g. locks you out, etc) that you are not to come to work for the next 2 weeks. The same applies to people resigning when they give notice (as the OP found out). HOWEVER, you are STILL CONSIDERED EMPLOYED by the company until the 2 week period is up. This is often important for Stock vesting, etc. And don't let them fool you that the "2 weeks pay" thing is a "severance package". It isn't. They are REQUIRED to pay you as long as you are an employee.
Don't Ever Sign Any End-Of-Employment Contract To Get a 2-Week Severance. You're an idiot if you do - they owe you the money in any case. The only time you should sign one of the agreements is to get money beyond what would be coming to you AFTER YOUR RESIGNATION/TERMINATION DATE (not the date you gave/received notice).
As a side note, this idiotic "walk-them-out-when-they-resign" policy seems to have originated in Silicon Valley in the 90s, as a consequence of the Dot-Com boom. Too many companies with no proper HR department not having any sort of a clue as to how to professionally hire/fire people. Unfortunately, it seems to have become a trend (it's the norm here in Silicon Valley for everyone, including the huge companies), which is telling as to the lowered quality of management (and HR) of companies these days.
-Erik
One of the biggest productivity saps for a sysadmin is dealing with the massive volume of email that we get. In even a moderate-sized business, it's easy to get 1,000 mails per day, with a couple 100 actually from a person, not an automated script.
Now, what I'm about to say is predicated on the assumption that your external mail server already runs SPAM filters, and that virtaully everything that you actually get is "real" mail. If this is not the case, FIX THIS FIRST. Get your company to pony up for some serious anti-spam software. It saves EVERYONE a ton of time, and at the same time, cuts down on your (the company's) exposure to the nasties that inhabit email.
First, pick an email client which has filters. My preference is for Evolution or Thunderbird, but there are many out there. Pick one. As a previous poster noted, GIVE IT ITS OWN DESKTOP - that is, in your window manager which has virtual desktops, dedicate one solely for the email client. Now, configure it with lots of filters to sort your mail. Personally, I have a reasonable hierarchy with 3 folders at the top level: NOW, LATER, and WHENEVER. Underneath these, there should be folders for every type of email you get: ones from your boss, ones from the company HR, ones from the monitoring scripts running on your servers (you do have these, right? RIGHT?). Take a good long time figuring out how to get these down cold - you want a good balance of sufficient sorting without going overboard. I find that having about 30-50 folders total is optimal for me. If you can, also have the email client tag your mail with "importance" color coding (most clients have this, and it's really useful).
Now to reading: obviously, your should read the NOW, well, NOW. However, you don't want to be completely interrupt-driven. I would turn off any biff-style mail notification, or at best, turn down its check time to no less than 10 minutes between check. Instead, train yourself to periodically check the NOW folder. Read and deal with the NOW stuff during your normal workflow.
The LATER folder should probably be read every couple of hours, or if you truly haven't anything else to do. Resist the temptation to open it and look. Finally, the WHENEVER shouldn't be read until the end of the day (or maybe while your eating lunch at your desk ;-)
Email is one of the great things about networks; however, it can be an enormous timesink if not properly handled. -Erik
I've noticed everyone here is recommending their favorite language, without mentioning why it might be useful for beginning programming. I'd like to address that issue:
In Beginning Programming, you're assuming that the student know absolutely nothing about programming. They should know Algebra (to understand variables), but that's about the extent of Math.
Thus, probably the Right Thing to do is build on the basic Math understanding that your students have, and start to introduce programming concepts from there.
For that reason, I'd start with a Functional Programming language as they tend to be rather obviously Math-derived, and help ease the introduction of programming constructs. I would recommend Scheme, since it is associated with one of the best "Teaching Programming" texts ever, Structure and Interpretation of Computer Programs.
After they've grasped Functional Programming, I'd have them move on to Procedural Programming, starting in Pascal or Modula-2. These two are excellent languages which illustrate the fundamentals of procedural programming, without some of the nastier pitfalls. There are also excellent textbooks available for Pascal (fewer for Modula-2). Later, I'd move them to C to introduce pointers and some of the other hairier features.
Finally, they're ready for Object Oriented Programming, for which I'd use Java - it's widely used, very common for college-level coursework, and there are a large number of supporting utilties and good textbooks useful for teaching aids.
Overall, I'd look at teaching Functional in 1 semester, Procedural in 1 or 2 semesters (depending on the detail and breadth you want), and OO in 1 or 2 semesters.
In all honestly, I love scripting languages like Perl, Python, and Ruby. However, as a teaching tool, they're all too multi-purpose, and it's easy for the student to do something they're not supposed to do (even though it works). For teaching languages, you want ones which pretty much only allow the student to program in the methodology you're teaching. That is, you generally want those languages which are LESS flexible, since your main goal is correctness, not functionality.
-Erik
Yes, Virginia, we can use mutli-core. I mean, we're all into SMP heavily in the non-desktop role (does anyone actually make a "server" that doesn't have SMP?)
There are two big things I love about the multi-core Opterons: They draw less power than equivalent SMP machines (acutally, quite a bit less), and they allow multiple "CPUs" to use the same memory controller. Nominally, the second isn't a big win, but it can be for practical purposes.
Opterons have dedicated memory channels on them, so a current dual-socket Opteron has two DISTINCT DIMM banks - that is, on a motherboard with 8 DIMM sockets, 4 are allocated to each CPU socket. So if you have only one CPU, you can only use 4 DIMM sockets. Since those 4 sockets are often configured as a single bank (i.e. they all have to be filled to work), you can't add another CPU to the system without buying more RAM. This is wasteful. But with a multi-core opteron, all on-chip cores share the same memory bank.
The jist of this is that it'll be easier to have High-Compute, lower RAM configurations than it currently is reasonable to do. There are a lot of tasks out there which it is really nice to have a modest amount of RAM (say 4GB), but with huge crunch. Currently, it's hard to buy a config to do that, since you generally either end up way over-paying for CPUs, a huge number of tiny DIMM chips (which sucks for future expansion), or a larger number of motherboards, which draws more power.
And, hey, they're not tooo bad in price. Sun's dual-core v40z is less than twice as expensive as their single-core v40z, and you save lots on power/cooling/space.
Overall, a nice win.
-Erik
I see a lot of people decrying the 9-year length of sentence as excessive. I'd like to promote the idea that its actually lenient, given the harm to society.
First, for those who haven't RTFA, this guy's crime wasn't just "spamming", it was the electronic equivalent of mail fraud. Take a look here for mail fraud penalties. Yup - that's right. Up to 5 YEARS per occurance. Not per person actually defrauded, but per mail sent.
Furthermore, we seem to want to punish "blue collar" crime (physical violence and theft) as somehow more heinous than "white collar" crime (usually fraud and theft of money or intangible property). As a poster above noted, blue collar crime tends to have a severe impact on a very limited number of people, though in the aggregate it also attacks a locality's social fabric (consider the number of violent crimes in someplace like South Central LA and the correlation to property values there). White collar crime, however, tends to impact a large number of people to a lesser extent, but also directly attacks the fundamental underpinnings of the society: in particular, the fundamental trust in fairness and shared responsibility that is essential for modern societies to function.
Fraud in particular is a particularly heinous crime from a societal standpoint, as it attacks the basic trust we put in financial transactions. A CEO giggering quarterly numbers is doing more than just cheating some stockholders out of a few cents in stock price - he's attacking the whole investing system which depends on truth in information dissemination. For if investors can't trust that a company's 10k annual report has real numbers, how can they invest?
White collar crime needs to be far more heavily punished than it currently is. And, it is much more deterred by increased prosecution and higher penalties than blue collar crime. Blue collar crime is generally only deterred by increased police presence (i.e. preventative measures) and not by increased penalties. White collar crime, on the other hand, generally shows a strong correlation to the likelihood of prosecution and severity of penalty. This is due to the fact that most white collar crime is committed by the more wealthy segment of the population, who generally do a risk analysis before committing the crime (i.e. "I'll steal $100,000 from the company, if I'm only 10% likely to get caught and only face 3 months in jail, but won't steal if I've a 50% chance of getting caught or if the sentence will be 5 years").
Also, remember that as "non-violent" criminals, white collar criminals tend to get put in low-security prisons, which cost much less to maintain than those in for violent crimes.
Overall, I'd like to see us start to put the emphasis on white collar crime instead of blue collar crime. In the big picture, I think it's far more damaging to society, and is far more frequent than people think.
There are some issues with this case (more specificly, the technicalities of the anti-spam law), but in the big picture, I think the sentence is exceedingly fair.
-Erik
Obviously, you've never met Thad's wife, Tavenner. She's fantastic in all senses of the word.
Hi Tavenner!
-Erik
Actually, it's a modified F-350 with a GE jet turbine.
More information
As cool as the DMC-based Time Machine is (and I have to admit, the original B-T-T-F movie is a good memory of my teenage years), the whole BB stuff just rocked. Soooo much more wacked, and so much more fun.
Now, if I can just get Kaneda's Bike from Akira...
-Erik
Honestly, if I were a games developer looking at the Linux market seriously, there is one feature which would really draw me in: the ability to provide a bootable distribution on the game CD.
One of the biggest headaches of game developers is trying to test their game on a sufficiently large subset of available hardware and software configurations to insure it will work properly. This isn't an issue on Consoles, which is one (not the only, but a big one) of the issues they are so popular to develop for. Having a bootable distro on the game CD gives the developer many of the advantages of both Console and PC:
Given the size of modern games, DVD distros are more likely than CD distros, but the concept is identical.
The bootable game CD/DVD has the potential to drastically reduce developer costs associated with modern games, and merge the best features of PC and Console gaming, with few drawbacks. I expect to see game makers venture into Linux in this area first.
-Erik
As an aside, 1 cubic meter of helium provides about 1 kg of lift (bouyancy). Hydrogen is only slightly better at about 1.2 kg per cubic meter.
So, if you want to hover a modest size craft (say 1 metric ton or so, which is rather small), you need to use about 1000 cubic meters of helium. Which is a bag 10 meters per size. Not a chance in hell for his design.
For example: a modern 737-800 aircraft weighs roughly 40 MT empty, has a cabin size of 40m long by 3.5m in diameter, and has a maximum cargo weight of about 30MT. Given that you can reduce the craft weight by 75% with advanced materials (very optomistically), you still need a gas volume of 40,000 cubic meters to lift the craft and cargo. This occupies a space 104 times the main cabin volume.
So, who is he trying to bullshit?
-Erik
Yes, he talks about a balloon-like structure. More specifically, a Zeppelin-like structure, with a rigid airframe holding the lighter-than-air gases. However, the design is of a plane-like glider. Take a look at a modern zeppelin. The radio of volume to surface area is many times greater than what he's proposing. Essentially, he's using lifting gases to enhance the performance of the craft. No matter what he's building the craft out of, it won't be able to hover on lift-gas alone, given his design.
The bottom line here is this: either you have to build a balloon/zeppelin-like craft with a large enough gas resevoir (in which case the aerodynamics of the structure completely preclude gliding), or you build a glider which uses wing lift to stay aloft. It's either or - you can't have both. The only way I could possibly think to provide enough lift for hovering in a glider is to have a hard vacuum instead of a helium (or hydrogen). But that would require a considerably stronger container, which in turn raises the total weight, Q.E.D.
He's designed an aeroplane, despite his attempts at misdirection.
-Erik
Honestly, after reading this site, it smacks of all the hype around the Segway (and many similar, less-successful "revolutionary breakthrough" schemes).
Yes, it is possible to create a fuel-less aeroplane that can still maintain forward motion. Advanced glider technology certainly fits this description. However, there are a couple of things that are missing from the adware:
The physics of underwater motion are similar those of flight - basic fluid dynamics here. The problem is he's ignoring fundamental environmental differences between water and air. The density differences between the two make it possible to move large masses underwater, but only tiny masses in the air using the same principles. Not to mention that the fluid consistency and motion between air and water are radically different, which invalidates using the ocean as a model for the sky.
He's a fraud. Pure and simple.
-Erik
Now, I usually don't get this annoyed, but we've gone through a bunch of these recently, and I'm sick of it.
The GPL is not Holy Scripture
The GPL is nice. It works for a large amount of stuff. However, it very much does NOT work for many other things, even in the Open Source world. I don't (and you shouldn't) want everything to be GPL'd. The GPL has a LOT of problems, freedom to copy aside. It is definitely not suitable for every purpose, given our current legal framework.
We should measure a license by how obnoxious and restrictive it is, not some idiot litmus test of GPL-compatibility. I prefer that we gravitate to a small number of general licenses for simplicity's sake, but there is no real good reason that they all HAVE to be GPL-compatible, any more than they all HAVE to be BSD-compatible.
The various ASL versions are all very benign and nice BSD-ish licenses, that may or may not be GPL compatible. They have very liberal code reuse and copying provisions, and very few restrictions. If they are GPL-incompatible, well, then, that's life. I'm not going to get angry over this, any more than I get upset because I can't use GPL libraries with my proprietary code.
Please quite focusing on the idiotic minutia, and pay attention to the hard issues of license lock-in and IP coralling prevelant in software licensing today.
-Erik
So far, the FCC and Congress haven't done much to regulate the VoIP and other technologies which will make VideoPhones ubiquitous in the coming years. (yeah, it's coming, whether we're all ready or not...)
I'd really rather NOT have them regulate this new area in the same manner which they regulate traditional telephony. However, I'm not in favor of the complete hands-off approach either. I'm hoping to see some sanity here - we certainly need a Universal Access Fee structure for Internet connections - if we want internet connections to be as common as telephones (and I think we can all agree that this is a Good Thing for the country), then we need to do a bit of subsizing.
Think of it this way: we plan to use Internet connectivity to Vote, use VideoPhone, standard Internet Telephony, and a host other things which we pretty much are going to designate as ESSENTIAL to living in the U.S. in a decade or two. Thus, we're going to have to subsize certain minimum connectivity offerings for those not able to afford the whole thing. Look in the front of your Telephone Book to see what I mean for current telephone service.
If we're lucky and a bit vigilant over the legislation being proposed, we can get the Right Thing done: basic Universal Service for everyone, while replacing the horrid mis-mash of telephony regulations with something much simpler and easier to deal with. If we're not, well, guess what - VideoPhone is going to be regulated even worse the regular telephony.
-Erik
A rather interesting article from the NYTimes covered some facts on a journalist's ability to protect sources:
Leaks and the Courts: There's Law, but Little Order (reg required)
I'd agree with the poster above that this was a "creative" use of the Carrier law by an FBI agent hoping to bully the guy into complying, with little actual legal ground to stand on. However, the reporter is definitely going to get a request for the full notes and info from the interview.
As the Times article points out, a reporter has no LEGAL means of keeping these away from the FBI, if really pressed to present them. We'll see if it gets that far.
-Erik
Section 4.3.1 of RFC 1034 pretty clearly states that the response to a name query is to be:
Now, the section thereafter goes on to talk about wildcards, so they are pretty much out of luck for saying that VeriSign isn't implementing the RFCs correctly. However, another portion of the RFC makes it very clear that wildcards are only for use within an entity's domain of control (that is, *.foo.com in DNS will not affect lookups under bar.com). The key here is that it is up to the OWNER of the domain in question as to the appropriateness of wildcards in DNS. VeriSign does NOT OWN THE .COM TLD. They merely ADMINISTER it for ICANN. Thus, there is a very good case for VeriSign being in breach of contract by failing to cary out the wishes of the OWNER of the .COM TLD. Which in this case is ICANN.
Basically, I would be a bit more thorough before going to VeriSign, but afterwards, I'd still wack them over the head with the contract and force them to remove the wildcard.
-Erik
I see a couple of people above moaning that SBC is forced to sell the "last-mile" loops to competitors for under cost, meaning the price they charge to the CLEC is less than it costs to maintain the wire.
This is patently false.
What the truth of the matter is this: the ILEC has to price the loop costs equally for all comers, INCLUDING THEIR OWN INTERNAL CUSTOMERS. Thus, out here in PacBell land, the costs that PacBell charges a CLEC (say AT&T or Covad) to lease a loop is the same costs is must charge PacBell Internet to lease that loop.
Guess, what? The ILEC like to subsidize their ISP and premium service groups by "selling" them loops for less than they cost. Regluation simply forces the ILEC to play fair, by allowing other CLECs to get this same price, and thus not allowing the monopoly on physical loop ownership to spill over into other services.
The ILEC could charge CLECs the proper amount to cover their costs, but they'd have to charge their in-house divisions the same rate. Thus, in reality, it is not the CLECs who are getting the free ride on the backs of the ILEC, but that the ILEC is propping up one of its own companies at the expense of another part of the ILEC conglomerate.
The Kansas legislature was completely correct - don't ever believe an ILEC "promise" in exchange for relaxation of regulation. They lie through their teeth constantly.
Fundamentally, the real solution is to force the ILECs to divest from physical loop ownership, and spin off a seperate company which is only allowed to own the loops, but may not sell data/voice services over those loops. Keep the hardware (a natural monopoly) distinct from the data (a natural competive market). Right now, we mix the two, to the detriment of all.
-Erik
First off, there is no need for the captain to be "instantly reachable". It's not like he's the only one which can make command decisions on a ship. It's been a while, but IIRC the title of the person who is in control of the ship is the Officer of the Deck. Should neither the captain nor Executive officer be on the bridge, one officer is designated the OOD and has effective command of the ship. Now, in a crisis, the XO and Captain almost always attempt to return to the bridge to reassert command, but the OOD can make all decisions (including breaking previous captain's orders, should the OOD deem it necessary) until relieved. So, it is silly to design a system to allow the captain to controll the ship from anywhere. Someone in the chain of command is already doing that from the place most suitable to do so, the bridge (or CIC, as appropriate).
Second, virtually all ships have a voice intercom systems set up throughout, which can relay orders back to the bridge far faster and more efficiently than some silly handheld WAP thingy. They're hardwired, so no emissions. They are invariably redundant, and far more likely to survive damage than a WAP system.
Finally, reduced manpower is a great goal, but generally is highly driven by putting in machinery which requires fewer operators. Communications systems are not really any manpower saver. And, as noted by others, you need twice as many people on a ship as it takes to operate all machinery: remember you have to run the ship 24x7, so you need at least two shifts (there's a little overlap, but 2x is a good rule of thumb), and you better have some extras for damage control and casualty replacement. So, you'll get manpower savings by automatic ammunition loading systems, better fire-supression, more efficient engines, better EW weapon systems, but not by adding WAP points.
Dumb idea.
-Erik
While they're technically a pure commertial venture, and thus I'm not exactly sure how closely they mimic the ClearChannel-style monopoly of "you-will-listen-to-this" music, I find that the BBC radio channels are an excellent alternative source of music.
BBC Radio Main Page
They have several HUNDRED streaming media files available for a huge variety of music, much of which is hard to find in the USA. I find it very refreshing, since British musical taste as always been a bit different than American taste, though the two tend to be quite compatible.
All their DJ shows are available on-line, and are updated weekly. And, since it's the BBC, they have virtually anything that you might be interested in, from Talk Shows about Scottish Sheepherding to the latest Dance club hits from the Continent.
-Erik
Actually, according to the FCC here in the USA, "Broadband" has a very specific meaning. That is, the download speed must be at least 208kbps. This is a legal definition. You can't sell a ISDN line and call it Broadband.
In general, that's a good definition from a networking standpoint, so I tend to stick with it in any case.
Oh, and the origin of the term "broadband" had NOTHING to do with the ability to send multiple signals on a single wire. It has to do with the SPECTRUM used to send those signals. Standard voice-only telephony uses a relatively small frequency range (about 4kHz). Broadband was originally used to describe an analog signal that used several multiples of this spectrum to send the data. Thus, it was much "broader" than the voice spectrum.
-Erik