Much the same way with the Wright Brothers. Something happened after their flight.
There are *many* stories like this throughout history. Whether or not the Wrights were the first to fly, it was thier flight that galvanized the world and led to sweeping change. Much as Christopher Columbus was almost certainly beaten to the American continent by centuries by other Europeans (notably the Vikings), his is nevertheless the voyage of discovery that is important, because it was the one that resulted in the New World becoming the focus of exploration, settlement, and trade for centuries.
FWIW, I also think it's a pretty iron-clad argument that the Wrights were the first to achieve any significant degree of *controlled* flight, however rudimentary. Their methodical and scientific approach to the problem through innovations such as wind-tunnel testing of hundreds of airfoils was clearly unique and the first application of real science to the problem of flying. Even the vaunted Dr. Samuel Langley of the Smithsonian was well behind in terms of basic research into the physical priciples of flight, which is perhaps one reason his steam-powered flying machine crashed into the Potomac only a short time before the Wright's first flight...
But its not for someone who could set up a Linux box.
Exactly. Which is why there are Linux distros like e-Smith and ClarkConnect that offer exactly the same sorts of functions to run on most any standard hardware.
This is nice, and interesting, but hardly newwsworthy, right? I mean, is there anything this box does that hasn't been done for years by other appliance servers from Cobalt, Gallantry, and the like? (Not to mention any ordinary machine loaded with one of the distros above...)
I wasn't aware that Dell was a good example of a potential big Sun customer.
Not just potential, actual. The big iron magic that runs Dell behind the scenes, especially their marketing/datawarehouse/OLAP systems were 100% Sun as recently as just a few years ago. How do I know? I was a Program Manager for Dell at the time and considered taking the job of running their "secret" Sun data center, where all the Sun gear went that was moved out of the glass customer showcase data center when it got too embarassing...
I expect that Dell has finally figured out how to wean themselves from Tandem and Sun now, but it had to be enormously expensive for them, especially on the Tandem/Compaq/HP side, which was even more vital than the Sun side since the Tandem was the heart of Dell's vaunted manufacturing process, and no one was left who even knew how the code worked, even for the parts of the source that they still had. (It was something goofy - a variant of Prolog or some other weird AI language...)
If I had a company with a janitor on the payroll....I'd have to pony up $100 to Sun for him because he's an employee? No thanks.
It's still *way* cheaper than any other alternatives (even the build-it-yourself option, in all liklihood, unless you're already a software company and can make a case for this sort of develpoment being a core competency (yeah, right...)), so paying a flat cost per employee is quite fair.
BTW, Jonathan Schwartz related a story during the launch event about the only customer he talked to that had serious reservations about the pricing: The Indian National Railroad, which has 15 MILLION employees - 1.5 Billion a year was a bit too rich for them... Perhaps, though, this is our golden chance to avenge all those jobs going to Bangalore...:-)
This is slashdot, people are supposed to mod when people say anything about charging for software. Everything should be free, we'll all suppose make our living doing support for the software.
[flame]No, actually, according to Stallman's stated position, we're supposed to all work for tips as waiters and give software away. It's not clear if even charging for support is permissible in the FSF's thinking, since that would involve putting money and software in the same context, something that apparently must never be done...[/flame]
Stop hiding behind the license-to-use crap and admit that software is a product. Make it do what it is advertised to do and what is described in the manual. If it has bugs, fix them at no cost to the customer. Then charge a fair price for the product and provide the support necessary for people to use it.
If you'd watched the lauch today, you would know that's pretty much exactly what Sun is proposing. Scott said as much directly when he pointed out that quality of most existing software is abysmal, and it costs an order of magnitude more than it should. The whole idea behind both the enterprise and desktop initiatives announced today is exactly the flat rate scenario you outline.
If Sun wishes to make money the very first thing they have to do is make a true and honest assessment on the state of the market and their place in it.
For all the brilliant minds that Sun has had under their roof over the years this is the one thing that they seem completely unable to do.
Believe me, they have. I just finished watching the live video of MacNealy and Scwhartz describing the new setup, and let me tell you - it is compelling.
Not only is it significantly cheaper than the Microsoft alternative, but Sun is rolling up absolutely all the network services enterprises really need into the deal - no piecemeal approach, no more wondering if this app server is compatible with that version of the mail backbone or whatever version of the operating system. Everything is integrated and tested to work together as a complete integrated, secure, and reliable system - desktops and servers, Linux or Solaris - it will just work. (And yes, I actually think they can pull this off - they already have most of the pieces in place, so it really just takes focus and execution. Focus is now a given - I'll reserve judgment on execution...)
This is the way we *should* have been doing IT for years, but no, we still spend way too much of our time doing system administration and in-house integration just to get the basics running.
Sun's approach will dramatically cut the cost of providing the basic infrastructure services that are the foundation for the interesting stuff. Garden variety systems administration *should* go away - and smart companies will see the potential to cast off the expensive shackles of today's current environment for a far more affordable and open future.
Anyone who believes in the strength of the network approach to delivering services can only applaud Sun's announcement today - I wish them the best in the market. This is an idea whose time has come, and it is definitely the way that Linux will start to make inroads against Microsoft in real-world business environments. Linux alone had no chance - Linux with Solaris, a complete set of services, all for a very affordable price, can hardly lose.
A lot of people here are suggesting using LDAP anyway, despite your statement that you don't have hte time or desire to learn al lthe arcana required to implement it.
In that case, then, I have a solution for you - the E-smith Linux distro, which will build you a dang useful server, including LDAP, within about 5 minutes of completing one of the easiest Linux installs you've ever encountered.
I don't bother with crap like configuring LDAP, Samba, firewalls, or mail servers anymore - E-smith has made it automatic and left me to add real value where it can do some good. Download an ISO and go to town - you'll be glad you gave it a try, and you'll have a complete working, properly configured LDAP server in mere minutes.
Entrepreneurs need capital though, generally, and they aren't going to be getting it from VCs in thsi climate:P.
No, it's not climate, it's structural - the real problem is that VC in this country has never been healthy and driven by entrepreneurism, but rather is structured to invest anywhere but in the companies that really should be invested in for the greater good of the economy. Companies that really need venture capital won't touch it with a ten foot pole, because the owners have too much of their life tied up in the company to give up the 80-90% required for a deal these days. That's bad for everyone - it means the people that need the funding don't get it, and the VCs continue to make investments to people who aren't really committed to making their companies work. This forces the VCs to play a sort of roulette, spreading the risk over many hopeless companies hoping one might be "the next Cisco" rather than making smaller, sustainable investments in companies that can show real, solid growth based on fundamental value.
Until this changes (and I don't see how, given the structure and collusion of US venture funds), the only option for most entrepreneurs is the same as it's always been: go into debt big time, and hope that it works out. (After making sure you have a decent chance of success, of course...)
And simotaniously(sic), the need for big boxes has decreased: its clusterd micros as far as the eye can see.
If you've ever really worked with big computing problems and clustered computing in the real world (I have, on the scale of 10,000+ deployed compute nodes in a single cluster) you know very well that clustered computing is powerful but certanly NOT a panacea, and that there are very significant types of big problems for which a cluster is pretty much useless, but a big SMP machine is ideal.
In general, Sun will have a tough time for another year or so, but will come out of this OK. That's at least partly because it's the easier problems that are suitable for clustered computing, and soon, people will want to move beyond those to the more challenging sorts of things that can't be easily ripped into neat little strips for the cluster.
For a general purpose unix workstation, a PC with Linux is cheeper, and more powerfull.. I daresay that the likes of Redhat is easier to manage then Solaris.
Suns are not the cheapest desktops, but can actually be pretty cost effective if you're serious about getting real work done rather than futzing with your work environment. There is no way Red Hat is easier to manage than Solaris - I've worked with both for years, and RH is still a major PITA - it's just that most of us have gotten so used to it that we forget how bad it aspirates. Each has it's strengths, but there are way too many people here on/. that apparently don't bother to learn the power of Solaris before they engage in politically (or karmally?) correct simultaneous Sun-bashing & Linux-fawning...
For entry level server, linux rocks.
Hmm, Sun thinks so too, which is why they have a line of low-cost Linux-based entry level servers (Cobalt)...
For mid range stuff, a cluster of linux boxen (sic) on Intel based SMP boxes is better then a single, or a smaller cluster of Suns.
Maybe, maybe not, for reasons mentioned above. Lots of us are eagerly awaiting to see if Sun is going to build an Opteron box that combines the power and flexibilty of that CPU with the awesome I/O and backplane interconnect of a classic Sun SMP server. 10 Gig (XG) Ethernet changes everything, and the problem with driving it isn't the interface, it's what you hook the interface to inside the box...
And for realy(sic) high end stuff, IBM is the only game in town: whatever else you can say about them they have made rock solid mainframes for 50 years, that work all the time, period. If you need such a machine, why would you risk getting one from a company that has been in that market for what? 2 years.
But mainframes are NOT Unix-like, open machines - they require very specialized proprietary knowledge to set up, run, and program. What you're missing here is that Sun is in the mainframe business, and has been for not two years, but at least 7, since the introduction of the E10000 Starfire, Sun's first mainframe-class machine. (Which actually started life as a Cray - Sun convinced SGI to sell this product off to them after SGI bought Cray. The E10K/E15K are serious machines in anyone's book, and it took IBM several years to figure out how to build a version of the SP2 that had equivalent backplane technology, in spite of the fact that the E10K's backplane technology was actually developed by another IBM division for Cray in the first place. Truth is sometimes stranger than fiction. At least it proves that IBM really can be an honest supplier of technology even to competitors without leaking info all over the company...)
Of course, you try to use Linux/BSD whenever possible so you don't ever have to see a $995 pricetag for a replacement keyboard. I bet that PCI card is hot swappable and well tested. It's the well tested that you are really paying for.
First of all, Sun's hardware is really pretty price competitive with serious gear from any vendor, even IBM or Dell. It's not the cheapest gear around, but it's affordable (unless you are too small to be serious about funding critical functions), but it just flat works, and will generally continue to work for so long that you start to wish it would die.
I've seen environments where I'm certain most of the Sun gear that was getting upgraded was subject to mass homicide - a prominent university here in Texas traded in a whole bunch of really old Suns they were using as X terminals - every one of them appeared to have had 120 VAC applied across the Ethernet circuitry. Heck, even Suns can't stand much of that...
Dell does that too. Of course the expense is amortized across a lot more customers.
Having worked for Dell and had the opportunity to go down to the call center and listen in on calls (long before they shipped the call center to a clueless corner of India), I can tell you that some fairly shockingly wrong and bad things are told to customers. I'm sure this happens elsewhere, too, but Dell is interested more in the customer's perception of the support process and the friendliness of the staff than they are in actually fixing the customer's problem.
Unfortunately, I've far too often seen tech support reps miserably fail to help the customer, but push all the right buttons to create a positive impression of th company in the customer's mind.
(Of course, I was stunned recently when I had a fairly obscure problem caused by a mix of BIOs and video card firmware incompatibilities, and Microsoft Tech Support helped me nail it in fairly short order - all in all, it was one of the best resolutions I've ever had to a really tough problem. Unfortunately, the info was in an unpublished knowledge base article, of which there are apparently quite a few - fat lotta good that does...)
The thing that really ticks me off is that at my school there are certain Sun keyboards that have the Ctrl key where you describe and others where that put it in the more expected place.
It can be a real bitch to get used to the configuration of the machine you happen to log into.
Actually, the control key next to the "a" is the expected place for any of us that predate PCs. This was the universally accepted location for the control key on nearly all terminals and workstations until the PC screwed it up. (Just like changing slashes to backslashes for no reason at all... ticks me off at least once or twice a week...)
The objects reside in a ring-shaped region called the Kuiper Belt, which houses a swarm of icy rocks that are leftover building blocks, or "planetesimals," from the solar system's creation.
This is nothing but pseudo-scientific drivel. These objects *may* be "leftover building blocks" from the formation of the Solar system, or they may not.
The fact is that we have no way of knowing. It's sloppy, unsubstantiated claims like this that (justifiably) undermine the credibility of those making such claims. Such bogus science consequently leads many of us to question the validity of the church of evolution on purely scientific grounds, if nothing else...
Although I haven't actually tried one myself yet (although I plan on ordering one later today), the SnapGear firewalls seem to be the most serious protection you can get at a really low price.
They're embedded Linux boxes, and SnapGear provides all upgrades for free, so there's no catch with having to have a maintenence contract as with some others.
They seem pretty serious (much better than the average consumer gear, but not much more money), have a nice set of capablities, and are far more reasonably priced than their more direct competition, which would be things such as SonicWall, SOHO or the like. There's even a very slick version built into an Ethernet NIC, ideal for double safety protection of the external interface of a firewall/gateway box like a PC loaded with the E-smith Linux distro. Yes, there are NIC drivers for this card for both Linux and Windows. I'm impressed enough to be looking at buying a couple to get solid and reliable VPN services without the hassles of doing that otherwise...
The most famous geek in the world was Milo Hoffman!
No, it would be the one and only Mr. Arlington Hewes of TPC, who lives on in all his virtuality at tpc.int, perhaps the last and only example of a bunch of Internet geeks getting a.int address...
Code generation is definitely something programmers of large/complex projects should look into. There's a lot of different forms of it, and I'd be surprised if people haven't used on form or another already.
One good example of how this can be applied in familiar terms is the excellent e-smith Linux server distro. E-smith uses a set of non-trivial Perl templates to generate all the server's configuration files, and another set of templates, based on FormMagick, for its server manager interface.
While it takes a bit to get fully up to speed on the templating system, (the downside to all code generation systems, but something that in this case is only required for programmers/developers, never users), it ensures that the correct thing always happens at the correct time. Once you realize that permanent changes must be made to the templates, and never the regular config files, as most admins are used to, it's easy.
The simplicity and robustness that this templating/code generation setup brings to the resulting system is why e-smith is the most solid and easy-to-use Linux *server* distro I've run across.
(Want a perfectly functioning firewall, mail server and Samba server (with Mac support) in minutes, without even having to know there's such a thing as *.conf files? E-smith is for you...)
There are other great examples of code generation in practice, but this is one that more people should be aware of...
Interesting ideas, but I have to question his use of the V2 as a model for mass rocketry. They were cheap, fast to build, and it only took a few people to launch one.
Probably because the Germans didn't have human cargo that they wanted to keep alive on top of them.
But that's not the point - if the Germans were able to build and launch *hundreds* of serious rockets a month at $13,000 each, all while getting the crap bombed out of them and having no labor to speak of, why can't we figure out a way to build a bunch of cheap rockets now, with 60 years of experience and technology on them?
Sure, man-rated gear will be far more costly, but it shouldn't cost thousands of times more, now should it?
* "rocket culture" at NASA
* "astronaut culture" at NASA
* materials science issues are quickly disappearing
* some probability of catastrophic (not deadly, just catastrophic) failure early on. must be budgeted using real-options analysis.
* 10-20B USD. This can easily be funded without "coalition" help. The US would soon own space like never before, as ESA's rockets would quickly look outdated.
* Defense concerns - the notion that a space elevator is vulnerable to, say, hostile fighter planes.
You forgot one of the most important roadblocks, one that sily pace elevator proponents always seem to gloss over: Just what would be the environmental effects of 100 miles worth of highly conductive carbon cable connecting the surface of the earth with the ionosphere? The effects of this on the earth's magnetic field are unknown, and possibly unknowable without experimenting with the only earth we've got.
I'm neither an environmentalist nor a conspiracy theorist, but the potential of such a structure to globally (and possibly permanently and significantly, according to the HAARP alarmists) affect the earth's environment are truly scary.
To reasoning people, such concerns will (and should) probably always prevent the construction of a space elevator.
(The whole idea is so goofy only the theoretical goofballs at MIT and their ilk that are so out of touch with the real world could have even suggested it. The space elevator could never be a real research program, but it could easily become another government boondoggle - a full employment program for theoretical technoids that can't cut it in the real world, but spin a mean bucket o' B.S. in PowerPoint...)
I already wrote a comment about this under the new launch vehicle topic, but it seems to be a better fit.
Those who haven't done so should read John Walker's (yep, the guy who wrote AutoCAD) paper written ten years ago on a different approach, one that *will* reduce the cost of spaceflight, and prove one way or the other if there is really enough commercial potential in space to build a sustainable space economy.
Note especially how there is valid historical documentation to support the viability of this aproach - it's not just blowing hot air, we have hard economic evidence that this both is doable and affordable.
It's time to kill NASA and do this right. What are we waiting for?
It's not too suprising that scientists are no match for millions of years of evolution.
Of course, there are lots of us that have valid scientific reasons for doubting evolution. I'll probably get modded into the gutter, but there are a number of very serious scientific problems with evolution, and most people don't even consider them, since that is heresy to the modern church of Evolution, and the indoctrination that passes for science education today.
There are many dozens more insightful articles in the Articles link above - even evolution supporters should read them if they want a balanced scientific view of the issue. The author of this site is not a dolt, but uber-programmer David Pogge (under the nom-de-plume Do-While Jones). Pogge is one of the few Fellows at the US Navy's China Lake Weapons Center, and is responsible for significant chunks of the US' dominance in missile guidance technology.
So, evolution is not an answer for everything, even the production of spider silk, and may not be an answer for anything, especially if one is interested in looking at the flimsiness of the "science" behind evolution.
It's worth pointing out that NASA is once again getting everything bass-ackwards and will once again try to use the highest (and therefore most expensive and unproven) technology possible to build it's future reusable spacecraft.
NASA has proven time and time again that it is unworthy of being entrusted with the US space effort and the enormous piles of taxpayer dollars it wastes through poor management and paralytic decision-making. (And yes, I've actually worked at JSC in Houston, so I speak with some insight.)
If space is really commercially viable (and that's a big if), there is a far better alternative that needs to be used, namely, the "Rocket a Day" approach. This was outlined by AutoCAD author John Walker in his paper titled "A Rocket a Day - Keeps the High Costs Awayten years ago now.
If you are even a little interested in an alternative way of looking at space travel, READ THIS PAPER, and pay special attention to the accomplishments of the Germans sixty years ago, which prove the argument.
A great read, and a direction that we should seriously investigate to replace the hopelessly outdated NASA juggernaut of waste, corruption, and buck-passing.
The DRM features I see actually described, in the article (did anyone actually bother to read it?) and through a couple of links in other posts above, are not terribly troublesome. Unfortunately, the paranoic responses here obscure that reality.
In fact, what MS has here is really nothing more than an MS implementation of a PGP equivalent with an authentication server and app hooks that know just a little bit about how the assigned rights relate to the nature of the files they're dealing with.
Let's take these paranoic rantings one at a time, shall we?
The critical presentation EXPIRES the night before you need it.
Could happen. You could also set the permissions on the file so that you can't read it. Why should this tool be a panacea to eliminate human stupidity? It won't be. This is a "power tool" - power tools can kill.
The only person with the rights to open a document is sick and didn't make the meeting.
Again, this one could happen, but if it does, it just shows that your group has woefully inadequate processes and procedures for assigning access to documents. (And how this is significantly different from that person having the copies in his briefcase and the source file in a non-open subdirectory on the server is not clear...) One presumes (since this is based on AD) that those higher in the auth chain could always override such assignments. In fact, this is one real benefit to a server-based auth method - it's effectively impossible to leave things locked up forever, or have rogue employees leave working time bombs.
The BIG customer tells you that they are not about to upgrade their servers and corporate software just to read your documents and tells you to provide material they can read or forget it.
This one would only happen if you were stupid enough to try to cram you auth methods down your customer's throats. You can do that with suppliers (although it's a losing move), but never customers. Generally, I expect this will not be widely used between companies, especially given the difficulties in establishing trust (both technically and humanly) between organizations.
They will have to have FULL-TIME rights managers, who track who is entitled to read whose documents.
You've never worked in the real world, have you? These people already exist, and have for better than 40 years, going back to NASA and the military-industrial complex. Their function is called "Configuration Managment" (do a Google search), and the idea is that these people determine what is kept, where, how, for how long, and who is allowed to use it in what ways. These are vital things any organization needs to do to manage information on a non-trivial scale, regardless of whether that work is building stealth bombers or growing organic kumquats.
And a full-time Search and Rescue team to retrieve lost documents, crack lost passwords, etc.
If it's any good, cracking will be fruitless. (I'll hold my judgment in reserve until I see MS implement real safety, but AD, for all its warts, has some really cool and elegant aspects. It's really too bad there aren't any interoperable alternatives.) As I mentioned above, the recursive nature of rights flow in AD (or any other decent auth system, such as Novell's Netware or NDS) should allow any employee's boss or other delegated person to override and/or reset rights.
In all, this is a decent solution to a very real problem. Unfortunately, it will certainly not be interoperable or standards-based, at least until and unless the Samba guys ever get a real interoperable AD replacement. And in big "enterprise" accounts, this will be a compelling feature that may persuade some customers to upgrade, something they are loathe to do now.
It's worth noting that there's no reason the same features couldn't be done in a completely open way if someone wante
Re:Curiously showing the size of apps & OSs
on
Windows 95 in 4.47MB
·
· Score: 1
>What I find telling as well is that the Mac OSX calculator.app is SIX times the size of the total RAM in the first Mac, and over twice the size of a complete OS install.
That's the "cruft" of a new software framework... it's a fact of computer life. The original Mac had 128 KB of RAM and a single internal 400 KB 3.5" floppy drive.
Remember, though, that the old Macs "cheated": Applications were tiny because most of the frequently-used routines lived in the Toolbox ROM, so that code was never part of the executable. It's an elegant and speedy system, but an architecture that has unfortunately fallen out of favor. Palm OS is perhaps the closest thing left today. Unfortunately, it looks like even that may be pushed out of the way by the bloatware approach over the next year or two...
Anyone who's tried to actually use CE or CE.Net in the real world has seen that it successfully combines all the worst features of small and large operating systems, with few of the benefits of either.
It's time for a lean OS for Embedded use - Linux has bloated to the point that it's hardly useful anymore for embedding - I can't even get any modern Linux to run on my 16 MB RAM Toshiba Libretto, which ran Win95 just fine. (It does run Mandrake 7.1, but it's pretty much stuck there...)
Slashdot Mirror
Much the same way with the Wright Brothers. Something happened after their flight.
There are *many* stories like this throughout history. Whether or not the Wrights were the first to fly, it was thier flight that galvanized the world and led to sweeping change. Much as Christopher Columbus was almost certainly beaten to the American continent by centuries by other Europeans (notably the Vikings), his is nevertheless the voyage of discovery that is important, because it was the one that resulted in the New World becoming the focus of exploration, settlement, and trade for centuries.
FWIW, I also think it's a pretty iron-clad argument that the Wrights were the first to achieve any significant degree of *controlled* flight, however rudimentary. Their methodical and scientific approach to the problem through innovations such as wind-tunnel testing of hundreds of airfoils was clearly unique and the first application of real science to the problem of flying. Even the vaunted Dr. Samuel Langley of the Smithsonian was well behind in terms of basic research into the physical priciples of flight, which is perhaps one reason his steam-powered flying machine crashed into the Potomac only a short time before the Wright's first flight...
But its not for someone who could set up a Linux box.
Exactly. Which is why there are Linux distros like e-Smith and ClarkConnect that offer exactly the same sorts of functions to run on most any standard hardware.
This is nice, and interesting, but hardly newwsworthy, right? I mean, is there anything this box does that hasn't been done for years by other appliance servers from Cobalt, Gallantry, and the like? (Not to mention any ordinary machine loaded with one of the distros above...)
I wasn't aware that Dell was a good example of a potential big Sun customer.
Not just potential, actual. The big iron magic that runs Dell behind the scenes, especially their marketing/datawarehouse/OLAP systems were 100% Sun as recently as just a few years ago. How do I know? I was a Program Manager for Dell at the time and considered taking the job of running their "secret" Sun data center, where all the Sun gear went that was moved out of the glass customer showcase data center when it got too embarassing...
I expect that Dell has finally figured out how to wean themselves from Tandem and Sun now, but it had to be enormously expensive for them, especially on the Tandem/Compaq/HP side, which was even more vital than the Sun side since the Tandem was the heart of Dell's vaunted manufacturing process, and no one was left who even knew how the code worked, even for the parts of the source that they still had. (It was something goofy - a variant of Prolog or some other weird AI language...)
If I had a company with a janitor on the payroll....I'd have to pony up $100 to Sun for him because he's an employee? No thanks.
:-)
It's still *way* cheaper than any other alternatives (even the build-it-yourself option, in all liklihood, unless you're already a software company and can make a case for this sort of develpoment being a core competency (yeah, right...)), so paying a flat cost per employee is quite fair.
BTW, Jonathan Schwartz related a story during the launch event about the only customer he talked to that had serious reservations about the pricing: The Indian National Railroad, which has 15 MILLION employees - 1.5 Billion a year was a bit too rich for them... Perhaps, though, this is our golden chance to avenge all those jobs going to Bangalore...
This is slashdot, people are supposed to mod when people say anything about charging for software. Everything should be free, we'll all suppose make our living doing support for the software.
[flame]No, actually, according to Stallman's stated position, we're supposed to all work for tips as waiters and give software away. It's not clear if even charging for support is permissible in the FSF's thinking, since that would involve putting money and software in the same context, something that apparently must never be done...[/flame]
Here's a new strategy:
Stop hiding behind the license-to-use crap and admit that software is a product. Make it do what it is advertised to do and what is described in the manual. If it has bugs, fix them at no cost to the customer. Then charge a fair price for the product and provide the support necessary for people to use it.
If you'd watched the lauch today, you would know that's pretty much exactly what Sun is proposing. Scott said as much directly when he pointed out that quality of most existing software is abysmal, and it costs an order of magnitude more than it should. The whole idea behind both the enterprise and desktop initiatives announced today is exactly the flat rate scenario you outline.
If Sun wishes to make money the very first thing they have to do is make a true and honest assessment on the state of the market and their place in it.
For all the brilliant minds that Sun has had under their roof over the years this is the one thing that they seem completely unable to do.
Believe me, they have. I just finished watching the live video of MacNealy and Scwhartz describing the new setup, and let me tell you - it is compelling.
Not only is it significantly cheaper than the Microsoft alternative, but Sun is rolling up absolutely all the network services enterprises really need into the deal - no piecemeal approach, no more wondering if this app server is compatible with that version of the mail backbone or whatever version of the operating system. Everything is integrated and tested to work together as a complete integrated, secure, and reliable system - desktops and servers, Linux or Solaris - it will just work. (And yes, I actually think they can pull this off - they already have most of the pieces in place, so it really just takes focus and execution. Focus is now a given - I'll reserve judgment on execution...)
This is the way we *should* have been doing IT for years, but no, we still spend way too much of our time doing system administration and in-house integration just to get the basics running.
Sun's approach will dramatically cut the cost of providing the basic infrastructure services that are the foundation for the interesting stuff. Garden variety systems administration *should* go away - and smart companies will see the potential to cast off the expensive shackles of today's current environment for a far more affordable and open future.
Anyone who believes in the strength of the network approach to delivering services can only applaud Sun's announcement today - I wish them the best in the market. This is an idea whose time has come, and it is definitely the way that Linux will start to make inroads against Microsoft in real-world business environments. Linux alone had no chance - Linux with Solaris, a complete set of services, all for a very affordable price, can hardly lose.
A lot of people here are suggesting using LDAP anyway, despite your statement that you don't have hte time or desire to learn al lthe arcana required to implement it.
In that case, then, I have a solution for you - the E-smith Linux distro, which will build you a dang useful server, including LDAP, within about 5 minutes of completing one of the easiest Linux installs you've ever encountered.
I don't bother with crap like configuring LDAP, Samba, firewalls, or mail servers anymore - E-smith has made it automatic and left me to add real value where it can do some good. Download an ISO and go to town - you'll be glad you gave it a try, and you'll have a complete working, properly configured LDAP server in mere minutes.
Entrepreneurs need capital though, generally, and they aren't going to be getting it from VCs in thsi climate :P.
No, it's not climate, it's structural - the real problem is that VC in this country has never been healthy and driven by entrepreneurism, but rather is structured to invest anywhere but in the companies that really should be invested in for the greater good of the economy. Companies that really need venture capital won't touch it with a ten foot pole, because the owners have too much of their life tied up in the company to give up the 80-90% required for a deal these days. That's bad for everyone - it means the people that need the funding don't get it, and the VCs continue to make investments to people who aren't really committed to making their companies work. This forces the VCs to play a sort of roulette, spreading the risk over many hopeless companies hoping one might be "the next Cisco" rather than making smaller, sustainable investments in companies that can show real, solid growth based on fundamental value.
Until this changes (and I don't see how, given the structure and collusion of US venture funds), the only option for most entrepreneurs is the same as it's always been: go into debt big time, and hope that it works out. (After making sure you have a decent chance of success, of course...)
And simotaniously(sic), the need for big boxes has decreased: its clusterd micros as far as the eye can see.
/. that apparently don't bother to learn the power of Solaris before they engage in politically (or karmally?) correct simultaneous Sun-bashing & Linux-fawning...
If you've ever really worked with big computing problems and clustered computing in the real world (I have, on the scale of 10,000+ deployed compute nodes in a single cluster) you know very well that clustered computing is powerful but certanly NOT a panacea, and that there are very significant types of big problems for which a cluster is pretty much useless, but a big SMP machine is ideal.
In general, Sun will have a tough time for another year or so, but will come out of this OK. That's at least partly because it's the easier problems that are suitable for clustered computing, and soon, people will want to move beyond those to the more challenging sorts of things that can't be easily ripped into neat little strips for the cluster.
For a general purpose unix workstation, a PC with Linux is cheeper, and more powerfull.. I daresay that the likes of Redhat is easier to manage then Solaris.
Suns are not the cheapest desktops, but can actually be pretty cost effective if you're serious about getting real work done rather than futzing with your work environment. There is no way Red Hat is easier to manage than Solaris - I've worked with both for years, and RH is still a major PITA - it's just that most of us have gotten so used to it that we forget how bad it aspirates. Each has it's strengths, but there are way too many people here on
For entry level server, linux rocks.
Hmm, Sun thinks so too, which is why they have a line of low-cost Linux-based entry level servers (Cobalt)...
For mid range stuff, a cluster of linux boxen (sic) on Intel based SMP boxes is better then a single, or a smaller cluster of Suns.
Maybe, maybe not, for reasons mentioned above. Lots of us are eagerly awaiting to see if Sun is going to build an Opteron box that combines the power and flexibilty of that CPU with the awesome I/O and backplane interconnect of a classic Sun SMP server. 10 Gig (XG) Ethernet changes everything, and the problem with driving it isn't the interface, it's what you hook the interface to inside the box...
And for realy(sic) high end stuff, IBM is the only game in town: whatever else you can say about them they have made rock solid mainframes for 50 years, that work all the time, period. If you need such a machine, why would you risk getting one from a company that has been in that market for what? 2 years.
But mainframes are NOT Unix-like, open machines - they require very specialized proprietary knowledge to set up, run, and program. What you're missing here is that Sun is in the mainframe business, and has been for not two years, but at least 7, since the introduction of the E10000 Starfire, Sun's first mainframe-class machine. (Which actually started life as a Cray - Sun convinced SGI to sell this product off to them after SGI bought Cray. The E10K/E15K are serious machines in anyone's book, and it took IBM several years to figure out how to build a version of the SP2 that had equivalent backplane technology, in spite of the fact that the E10K's backplane technology was actually developed by another IBM division for Cray in the first place. Truth is sometimes stranger than fiction. At least it proves that IBM really can be an honest supplier of technology even to competitors without leaking info all over the company...)
Of course, you try to use Linux/BSD whenever possible so you don't ever have to see a $995 pricetag for a replacement keyboard. I bet that PCI card is hot swappable and well tested. It's the well tested that you are really paying for.
First of all, Sun's hardware is really pretty price competitive with serious gear from any vendor, even IBM or Dell. It's not the cheapest gear around, but it's affordable (unless you are too small to be serious about funding critical functions), but it just flat works, and will generally continue to work for so long that you start to wish it would die.
I've seen environments where I'm certain most of the Sun gear that was getting upgraded was subject to mass homicide - a prominent university here in Texas traded in a whole bunch of really old Suns they were using as X terminals - every one of them appeared to have had 120 VAC applied across the Ethernet circuitry. Heck, even Suns can't stand much of that...
Dell does that too. Of course the expense is amortized across a lot more customers.
Having worked for Dell and had the opportunity to go down to the call center and listen in on calls (long before they shipped the call center to a clueless corner of India), I can tell you that some fairly shockingly wrong and bad things are told to customers. I'm sure this happens elsewhere, too, but Dell is interested more in the customer's perception of the support process and the friendliness of the staff than they are in actually fixing the customer's problem.
Unfortunately, I've far too often seen tech support reps miserably fail to help the customer, but push all the right buttons to create a positive impression of th company in the customer's mind.
(Of course, I was stunned recently when I had a fairly obscure problem caused by a mix of BIOs and video card firmware incompatibilities, and Microsoft Tech Support helped me nail it in fairly short order - all in all, it was one of the best resolutions I've ever had to a really tough problem. Unfortunately, the info was in an unpublished knowledge base article, of which there are apparently quite a few - fat lotta good that does...)
The thing that really ticks me off is that at my school there are certain Sun keyboards that have the Ctrl key where you describe and others where that put it in the more expected place.
It can be a real bitch to get used to the configuration of the machine you happen to log into.
Actually, the control key next to the "a" is the expected place for any of us that predate PCs. This was the universally accepted location for the control key on nearly all terminals and workstations until the PC screwed it up. (Just like changing slashes to backslashes for no reason at all... ticks me off at least once or twice a week...)
The objects reside in a ring-shaped region called the Kuiper Belt, which houses a swarm of icy rocks that are leftover building blocks, or "planetesimals," from the solar system's creation.
This is nothing but pseudo-scientific drivel. These objects *may* be "leftover building blocks" from the formation of the Solar system, or they may not.
The fact is that we have no way of knowing. It's sloppy, unsubstantiated claims like this that (justifiably) undermine the credibility of those making such claims. Such bogus science consequently leads many of us to question the validity of the church of evolution on purely scientific grounds, if nothing else...
Although I haven't actually tried one myself yet (although I plan on ordering one later today), the SnapGear firewalls seem to be the most serious protection you can get at a really low price.
They're embedded Linux boxes, and SnapGear provides all upgrades for free, so there's no catch with having to have a maintenence contract as with some others.
They seem pretty serious (much better than the average consumer gear, but not much more money), have a nice set of capablities, and are far more reasonably priced than their more direct competition, which would be things such as SonicWall, SOHO or the like. There's even a very slick version built into an Ethernet NIC, ideal for double safety protection of the external interface of a firewall/gateway box like a PC loaded with the E-smith Linux distro. Yes, there are NIC drivers for this card for both Linux and Windows. I'm impressed enough to be looking at buying a couple to get solid and reliable VPN services without the hassles of doing that otherwise...
The most famous geek in the world was Milo Hoffman!
.int address...
No, it would be the one and only Mr. Arlington Hewes of TPC, who lives on in all his virtuality at tpc.int, perhaps the last and only example of a bunch of Internet geeks getting a
Code generation is definitely something programmers of large/complex projects should look into. There's a lot of different forms of it, and I'd be surprised if people haven't used on form or another already.
One good example of how this can be applied in familiar terms is the excellent e-smith Linux server distro. E-smith uses a set of non-trivial Perl templates to generate all the server's configuration files, and another set of templates, based on FormMagick, for its server manager interface.
While it takes a bit to get fully up to speed on the templating system, (the downside to all code generation systems, but something that in this case is only required for programmers/developers, never users), it ensures that the correct thing always happens at the correct time. Once you realize that permanent changes must be made to the templates, and never the regular config files, as most admins are used to, it's easy.
The simplicity and robustness that this templating/code generation setup brings to the resulting system is why e-smith is the most solid and easy-to-use Linux *server* distro I've run across.
(Want a perfectly functioning firewall, mail server and Samba server (with Mac support) in minutes, without even having to know there's such a thing as *.conf files? E-smith is for you...)
There are other great examples of code generation in practice, but this is one that more people should be aware of...
Actually, Oxygen *IS* blue - If you don't believe me, have a look at it as you pour it out in liquid form from a dewar.
You'll find it's a familiar light blue, perhaps even sky blue...
Interesting ideas, but I have to question his use of the V2 as a model for mass rocketry. They were cheap, fast to build, and it only took a few people to launch one.
Probably because the Germans didn't have human cargo that they wanted to keep alive on top of them.
But that's not the point - if the Germans were able to build and launch *hundreds* of serious rockets a month at $13,000 each, all while getting the crap bombed out of them and having no labor to speak of, why can't we figure out a way to build a bunch of cheap rockets now, with 60 years of experience and technology on them?
Sure, man-rated gear will be far more costly, but it shouldn't cost thousands of times more, now should it?
Roadblocks:
* "rocket culture" at NASA
* "astronaut culture" at NASA
* materials science issues are quickly disappearing
* some probability of catastrophic (not deadly, just catastrophic) failure early on. must be budgeted using real-options analysis.
* 10-20B USD. This can easily be funded without "coalition" help. The US would soon own space like never before, as ESA's rockets would quickly look outdated.
* Defense concerns - the notion that a space elevator is vulnerable to, say, hostile fighter planes.
You forgot one of the most important roadblocks, one that sily pace elevator proponents always seem to gloss over: Just what would be the environmental effects of 100 miles worth of highly conductive carbon cable connecting the surface of the earth with the ionosphere? The effects of this on the earth's magnetic field are unknown, and possibly unknowable without experimenting with the only earth we've got.
I'm neither an environmentalist nor a conspiracy theorist, but the potential of such a structure to globally (and possibly permanently and significantly, according to the HAARP alarmists) affect the earth's environment are truly scary.
To reasoning people, such concerns will (and should) probably always prevent the construction of a space elevator.
(The whole idea is so goofy only the theoretical goofballs at MIT and their ilk that are so out of touch with the real world could have even suggested it. The space elevator could never be a real research program, but it could easily become another government boondoggle - a full employment program for theoretical technoids that can't cut it in the real world, but spin a mean bucket o' B.S. in PowerPoint...)
I already wrote a comment about this under the new launch vehicle topic, but it seems to be a better fit.
Those who haven't done so should read John Walker's (yep, the guy who wrote AutoCAD) paper written ten years ago on a different approach, one that *will* reduce the cost of spaceflight, and prove one way or the other if there is really enough commercial potential in space to build a sustainable space economy.
Here's the link to the paper: A Rocket a Day - Keeps the High Costs Away
Note especially how there is valid historical documentation to support the viability of this aproach - it's not just blowing hot air, we have hard economic evidence that this both is doable and affordable.
It's time to kill NASA and do this right. What are we waiting for?
It's not too suprising that scientists are no match for millions of years of evolution.
Of course, there are lots of us that have valid scientific reasons for doubting evolution. I'll probably get modded into the gutter, but there are a number of very serious scientific problems with evolution, and most people don't even consider them, since that is heresy to the modern church of Evolution, and the indoctrination that passes for science education today.
Those that have more open minds, and are willing to explore *why* science actually militates against evolution are invited to read the articles at Do-While Jones' excellent Science Against Evolution site.
I particularly recommend the following as a getting started place:
Let's talk about Lucy - Things you never knew about the famous invented ape
Radioactive Dating explained, Part 1 - A real eye-opener
Radioactive Dating Explained - Part 2
The Species problem - how DNA studies do NOT support evolutionary theory
There are many dozens more insightful articles in the Articles link above - even evolution supporters should read them if they want a balanced scientific view of the issue. The author of this site is not a dolt, but uber-programmer David Pogge (under the nom-de-plume Do-While Jones). Pogge is one of the few Fellows at the US Navy's China Lake Weapons Center, and is responsible for significant chunks of the US' dominance in missile guidance technology.
So, evolution is not an answer for everything, even the production of spider silk, and may not be an answer for anything, especially if one is interested in looking at the flimsiness of the "science" behind evolution.
It's worth pointing out that NASA is once again getting everything bass-ackwards and will once again try to use the highest (and therefore most expensive and unproven) technology possible to build it's future reusable spacecraft.
NASA has proven time and time again that it is unworthy of being entrusted with the US space effort and the enormous piles of taxpayer dollars it wastes through poor management and paralytic decision-making. (And yes, I've actually worked at JSC in Houston, so I speak with some insight.)
If space is really commercially viable (and that's a big if), there is a far better alternative that needs to be used, namely, the "Rocket a Day" approach. This was outlined by AutoCAD author John Walker in his paper titled "A Rocket a Day - Keeps the High Costs Away ten years ago now.
If you are even a little interested in an alternative way of looking at space travel, READ THIS PAPER, and pay special attention to the accomplishments of the Germans sixty years ago, which prove the argument.
A great read, and a direction that we should seriously investigate to replace the hopelessly outdated NASA juggernaut of waste, corruption, and buck-passing.
The DRM features I see actually described, in the article (did anyone actually bother to read it?) and through a couple of links in other posts above, are not terribly troublesome. Unfortunately, the paranoic responses here obscure that reality.
In fact, what MS has here is really nothing more than an MS implementation of a PGP equivalent with an authentication server and app hooks that know just a little bit about how the assigned rights relate to the nature of the files they're dealing with.
Let's take these paranoic rantings one at a time, shall we?
The critical presentation EXPIRES the night before you need it.
Could happen. You could also set the permissions on the file so that you can't read it. Why should this tool be a panacea to eliminate human stupidity? It won't be. This is a "power tool" - power tools can kill.
The only person with the rights to open a document is sick and didn't make the meeting.
Again, this one could happen, but if it does, it just shows that your group has woefully inadequate processes and procedures for assigning access to documents. (And how this is significantly different from that person having the copies in his briefcase and the source file in a non-open subdirectory on the server is not clear...) One presumes (since this is based on AD) that those higher in the auth chain could always override such assignments. In fact, this is one real benefit to a server-based auth method - it's effectively impossible to leave things locked up forever, or have rogue employees leave working time bombs.
The BIG customer tells you that they are not about to upgrade their servers and corporate software just to read your documents and tells you to provide material they can read or forget it.
This one would only happen if you were stupid enough to try to cram you auth methods down your customer's throats. You can do that with suppliers (although it's a losing move), but never customers. Generally, I expect this will not be widely used between companies, especially given the difficulties in establishing trust (both technically and humanly) between organizations.
They will have to have FULL-TIME rights managers, who track who is entitled to read whose documents.
You've never worked in the real world, have you? These people already exist, and have for better than 40 years, going back to NASA and the military-industrial complex. Their function is called "Configuration Managment" (do a Google search), and the idea is that these people determine what is kept, where, how, for how long, and who is allowed to use it in what ways. These are vital things any organization needs to do to manage information on a non-trivial scale, regardless of whether that work is building stealth bombers or growing organic kumquats.
And a full-time Search and Rescue team to retrieve lost documents, crack lost passwords, etc.
If it's any good, cracking will be fruitless. (I'll hold my judgment in reserve until I see MS implement real safety, but AD, for all its warts, has some really cool and elegant aspects. It's really too bad there aren't any interoperable alternatives.) As I mentioned above, the recursive nature of rights flow in AD (or any other decent auth system, such as Novell's Netware or NDS) should allow any employee's boss or other delegated person to override and/or reset rights.
In all, this is a decent solution to a very real problem. Unfortunately, it will certainly not be interoperable or standards-based, at least until and unless the Samba guys ever get a real interoperable AD replacement. And in big "enterprise" accounts, this will be a compelling feature that may persuade some customers to upgrade, something they are loathe to do now.
It's worth noting that there's no reason the same features couldn't be done in a completely open way if someone wante
>What I find telling as well is that the Mac OSX calculator.app is SIX times the size of the total RAM in the first Mac, and over twice the size of a complete OS install.
That's the "cruft" of a new software framework... it's a fact of computer life. The original Mac had 128 KB of RAM and a single internal 400 KB 3.5" floppy drive.
Remember, though, that the old Macs "cheated": Applications were tiny because most of the frequently-used routines lived in the Toolbox ROM, so that code was never part of the executable. It's an elegant and speedy system, but an architecture that has unfortunately fallen out of favor. Palm OS is perhaps the closest thing left today. Unfortunately, it looks like even that may be pushed out of the way by the bloatware approach over the next year or two...
Anyone who's tried to actually use CE or CE.Net in the real world has seen that it successfully combines all the worst features of small and large operating systems, with few of the benefits of either.
It's time for a lean OS for Embedded use - Linux has bloated to the point that it's hardly useful anymore for embedding - I can't even get any modern Linux to run on my 16 MB RAM Toshiba Libretto, which ran Win95 just fine. (It does run Mandrake 7.1, but it's pretty much stuck there...)