Instead of creating meta-machines that serve a lot of users, home and otherwise, he's correctly pointed out that machine makers aren't serving other markets, like having cogent product line including mobiles, PDAs worth a crap, and so on. Computer makers don't understand the CE marketplace at all, and so it's no wonder they can't make money at it.
Michael Dell was the first to figure it out, but others have had moderate success driving consumer features. What they can't drive is consumer operating systems, because there ARE NONE.
isochonous applications are here to stay, including real-time video, lots of VoIP, and others whose time-domain is sensitive to latencies. The expertise of the writer involved is suspect, quoting Son-- who ran Softbank and isn't an engineer. Routing protocols, QoS, route saturations, re-authentication cycles, all of these cause objectionable latencies-- not to mention the end-2-end capacity of the network involved.
Yes Martha, we need bandwidth. Ignore the idiot.
Case: email
Case: IM
Case: online gaming
Case: forums
Case: surfing
Case: RSS
Case: interactive purchasing
Case: downloading entertainment
Case: blogging
Case: social infrastructure
Go on. Run the sieve. Tell me what's not addictive. We're social and interactive creatures. Ask the question again. What knid of dumbass question is this? Yo: Cowboy Neal--> learn to ask a reasonable question.
No load charity is a wonderful thing. The current model of huge organizations with powerful overheads taxes donations. If you consider the actions of the administrators of charity, they're much like a program in numerous ways.
I'm not saying that eliminating administration can be perfected (you need audits, and program adjustments along the way), but a higher efficiency model would be welcomed. Bureaucracies tend to grow until they suck away the intent of charitable organizations. Some are far more efficient than others.... but in a future world perhaps more of contributions gets to the need, rather than the machine that services the need.
Bad article: 2 diff things for 2 diff purposes
on
Java Is So 90s
·
· Score: 1
The whole Java hegenomy is a different culture for different apps than LAMP/SAMP etc. "So 90's...." doesn't talk well about platform compatibility, resource management, coding methodologies, and numerous other characteristics. It's like comparing an airbrush to an airchisel. Ok, bad metaphor but also a bad comparison.
Maybe the correct phrase is a hegenomy of devices, as this is what we have.
To extend your checking flights metaphor, I can do this on my mobile, my PDA, my notebook, or a terminal somewhere I don't own (not that I would). Each device is running something different. The mobile runs Symbian; the PDA runs WinCE, the notebook runs MacOS, and only heaven knows what the public terminal has, probably a Windows session.
At the core on the thick side is (statistically, anyway) either Apache/Tomcat, or IIS/something running the back end. Maybe Solaris, HP/UX, or something else is behind the curtain. Sun is trying to sell what's behind the curtain without thinking about the rest of the capability of the delivery system or the end device. Indeed the end device should go away or become something very uniform and manageable by their last perceived closed app, Java.
Yuck. I don't think that behind the curtain model works. Yes, hulking fast servers are good things. But divorcing what's at the edge is really very silly, unless you're a hardware server maker like Sun--- who provides none of those edge devices-- so in their minds they must not exist. These devices aren't embraced, they're ignored. It's egalitarianism through ignorance and hubris.
This is the same thick model they've been bandying about since inception, and failing-- except during the dot-bomb era when people just bought hardware for mindless reasons and irrational exuberance. As Robert Plant might sing, the song remains the same.... just a new stanza.
Sun is otherwise pretty smart, and smarter than Red Hat and SuSE when it comes to Unix. But they're also stuck in their own mud. McNealy and Schwartz should exit, and get a team that can appeal to a new and differently incented group of buyers. But their egos get in the way. They always do.
1) AT&T and Western Electric were a monopoly, as in NO ONE ELSE GETS TO PLAY. CPE, as in phones, PBXs, and so on, were 100% controlled until PBXs broke loose, then the Green Decision to break them up. Their various groups were split into pieces that followed business unit logic: Lucent, AT&T Long Lines, and so on. History marks this
Time Warner, on the other hand, is a media empire with broadband holdings. In some areas, BrightHouse and other subsidiary divisions provide entertainment/TV Cable, and adjunct to that, broadband access. Their penetration is good in the US, and the US only. AOL was once the king of dial-up. No more. Dialup isn't a good situation. But as you can live without a phone, you can live without broadband access. Cable TV is like a utility, and uses utility metaphor, and has been promoted legally to the sense of a utility in a handful of markets.
But TW has serious competition in each business segment. Comcast, Adelphia, not to mention Verizon, SBC, BellSouth, and so on. AOL has competition: MSN, Google, Yahoo, and numerous others. TW publications have competition across the board. Their publishing unit has excellent and cogent competition across the board. THIS IS NOT A MONOPOLY.
2) I applaud the breakup of TW for numerous reasons. It will make Brighthouse stronger, and not beholden to the rotten cash problems that TW has. TW has demonstrated time and again that they're not leaders, and barely followers in this area. AOL needs fresh blood and new ideas. Their business model made sense when broadband build-out was small and spotty. It's not anymore. The added value needed for AOL would be to buy something like MySpace and add put it on the google revenue model. But AOL is leaden, and Steve Case is gone, and I think personally that they'll slowly lose share until they croak or are sold off for assets.
3) There are no other parallels that exist in the comparison. It's ok that to say, "gosh, bad idea". I have to do this, too. If you want to have a discussion on the problems of national broadband inconsistencies, or how publishing content and revenue models now have become divorced from providers, or how IPTV will disrupt the living hell out of everything, then lead on. But it's an apples vs durian discussion here. They might be both businesses, but it really stops there. And like durian, AT&T stank.
4) Competition? Every telco. Every cable company. Every magazine/book publisher. Every content generator. There is no current problem with broadband competition. There may be as, like the recent 2600 Magazine cover implies, a new Bell, as they seem to re-unite again. But the telcos are only now putting in fiber, and worse, passive fiber. The last mile is still an open book, and the chapters are by no means written yet.
So, no. GIGO. SHOUTING IN CAPITALS DOESN'T INCREASE THE COMPETENCE OF A POST.
This goes further into the model proposed by the post: 8-cores in the CPU (one FPU for the eight) and 32 discrete threads, all in a 2U server box. This is based on UltraSparc, but there's Solaris 10, and the port of gcc to it for seductive app transfer. The whole idea is a hardware play.
It makes me wonder why there must always be this gulf between hardware and software vendors. The most successful models meld them together handsomely into devices like iPods, mobile/pda devices, etc. This thick-thin shift is so insane. At the end of the day, we just want to do work, entertainment, and something useful with the devices we buy, and the location of what's going on is increasingly irrelevant. But perhpas this is what (F)OSS software will get for us, an army of coders coupled to an army of blade vendors, with dumb devices at the edge.
Ma Bell, when she was an aggregate whole, had over $4B in assets in 1972 related to ROW, easements, and property that was a sanctioned utiility monopoly.
Time Warner has assets, that in 2005 dollars, are but a fraction. Indeed most of their distribution infrasturcture rides other utility ROW and easements. They are by no means a monopoly, although they do enjoy franchise status in numerous areas. That franchise can be revoked if the company is re-capitalized. Federal law also supports numerous re-examination possibilities when they recapitalize, or alter their structure, or are merged. It's non-trivial.
Equating it to the break of Ma Bell just doesn't stick to the wall. The issues are just not the same. I analyze the comm/telecom industry deeply. Regretfully for the poster, they aren't equivalent, not even slightly.
Comparisons to the Ma Bell breakup don't work here; that's the point.
Icahn is trying to up his asset. Ma Bell was a utility, and a monopoly. Time Warner is neither.
Time Warner has diverse media and manufacturing assets, and while they have content generation and distribution, they're not a monoply in any of the areas of their business, like Ma Bell was.
A breakup in this case, has no useful metaphor to Judge Greene's breakup of the AT&T components into different companies. The only common denominator between the two, is that the fact that they were broken up to increase shareholder value. Other comparisons aren't valid. This is an economic issue, not one with parallels.
Comparisons that try to link the two in that way just don't work-- TW's asset structure, business model, governmental regulatory control, and every other facet between TW and AT&T are different. Really different. The supposition of the post tries to presume that tie, where no tie exists.
There is no dependent public service on Time Warner. All they things they do, especially in media, can't be considered in the same way that the monopoly of AT&T and their FCC and PUC/state-governed telephony was.
These two aren't equal; they aren't congruent, they aren't even parallel.
Time Warner's broadband properties are not a utility, like water, sewer, electricity, or natural gas-- things you can't live without. They don't have the same history, the same economics, the same monopoly control, the same easement and right-of-way capital assets, and so on.
Therefore, garbage-in, garbage-out. The comparison is null, and it is, unfortunately moot.
There are people with spinal cord and brain injuries that could really benefit from this. People with trauma recovery with no sensation that don't realize an appendage is bumping into something. It's not quite like having skin with nerve endings, rather it's an early alert that senses something that you either can't, or can't *yet*.
In another way, it's also a way to help people recover from muscular atrophy, sensing leg movements, or arm movements. It can tell you when something's too tight, or incorrectly applied. Think physical therapy, or improving your golf swing, football kick, or reducing RSI.
Although I don't understand its resolution capability, it could also be used for carpet-fabric that could tell people when someone's at the door, or that someone has been in a room, or that the person weighs 100kg, etc.
Use your imagination beyond sex. I find this fairly fascinating.
This isn't a new or different idea at all
on
Internet Immunization
·
· Score: 2, Insightful
There are already appliance makers that do this very thing: identify malware and viruses, and signal the others, usually in the guise of spam control appliances.
Webs of early notifiers is also not a new idea; look at the honeypot networks that are on the web, the honeypot project, and so on.
The containment cited is theoretical, subject to the ability to correctly identify behavior, and doesn't prevent users from clicking on URLs that have malware, or filter signatures that have fast breakout behavior.
And so, the merit of the Nature article is in question. It's just a PR release in disguise.
My observation is that you're a coder. Few others write such long sentences.
There are cogent arguments supporting micro- vs traditional kernel designs. Microkernels are valuable when you control hardware architectures, or where one goes heavily to assembler to make a CPU sing. They're less fragile, but also mandate are more serious hardware examination approach to processor optimizations. That's why Apple can do it but it's tougher for others. Their jump to Intel proves that the Darwin kernel approach first perfected in Mach is actually not as tough to remake onto other platforms. I've been speculating that the OpenDarwin tree was more than just a curiosity on Intel/AMD platforms. It's the Trusted Computing area that pushed the decision over the edge, in my belief. Jobs has foud the mirth of making Hollywood happy. Jobs moved the Carnegie Mellon work done on Mach to Apple to be manifested under the Apple design methodology, and did so successfully after a few mild, non-showstopping mods. The result has been positive. Adding iTunes was astute.... four times he's done this, first with the Apple ][, the Mac, the MacOS, and now the *Pod designs. But it isn't for all, and simply points to the successes of an avatar.
Solaris/UltraSparc could take advantage of the same model... but they also own keys to Unix and BSD-- hard fought and won and politically continue the SunOS legacies in Solaris/Open Solaris.
Longhorn will be interesting to see. At some point, Microsoft must balance legacy architectural mistakes made in the fast-and-loose era and hone a better core architecture. This will be tough for them. But it'll be less expensive than the madness they face now. Their security models have improved vastly, and the raw brainpower purchased with Connectix, ZoomIt and others is paying handsomely. But it's closed, and bereft of numerous allegiances that power (F)OSS. The concept of free rocks them, and Google's kinda-free methods are infecting them. It'll be interesting to see what happens in five years.
The DOS->Win31->win95-win98-winME family is ideologically and in reality built on DOS. This model wasn't ever a good one for multi-user anything, and was constrained to hybrids of the peer model rather than LAN/WAN hierarchical resource utilization/RPC models.
The OS/2->Winnt31/35->NT4->NT5/Win2K/XP/Win2003 is a different family and was better, hobbled by a legacy addiction to the IBM legacy models of OS/2 and the PDC/BDC/LAN Manager legacy. Windows 2000 evolved the Active Directory concept that put still another nail in Novell's coffin (please don't go to NetWare vs AD, etc.).
The end of the story says that you're right, but partly for different reasons than you state, IMHO. Microsoft did evolve object classes and API sets that are pretty clever. But when so many are strung together, and become dependent, it causes enormous delta in what can be easily changed and evolved. That's why new products and releases are so leaden in getting through the dev process. This is also coupled to the fact that other developer partners are loathe to spend money to constantly update code-- and this is the motivation behind Microsoft's stated goal of major releases every four years with revised kernels within the same release every two years. Look for "R2" of Win2003 server coming VERY soon.
Linux, by contrast, is done when its done. It views collective substrate modifications at different levels.It also evolves older editions of the kernel (e.g. the 2.2, 2.4 tree development) as active projects and maintenance obligations.
The Darwin kernel that underlays MacOS is different still, but is a microkernel approach and requires even less dependency checks because once a port is done, the compiled kernel fixes rarely have an effect on core API sets and distributed (mostly BSD) applications. Apple controls its developer network much better, and developers mix BSD, GNU, and other OSS and OSS-like models in a blend/blur.... some open some closed. But Apple carefully controls the microkernel, and so far, microkernel patches in the past four years have been surprisingly few, and Apple's problems equally few from a security/functionality patch perspective.
Solaris has similar qualiities, and now that gcc works well on Solaris, ports of various apps have found a new home on a stable environment.
The open design of Open Solaris, many BSD branches, and Linux have various benefits here. And, as seen in another reply, occasional lapses based on simple motivation to fix bugs.
Collaboration is locationally independent. Lots of on-line whiteboards exist, if and where needed. The lingua franca is English, and C, maybe with some perl and php thrown in.
I like EU coders, Balkan coders, Indo-Pakastani coders, Chinese coders, Aussies, all of them everywhere. Coding is a pretty personal exercise, even in teams. Collaborative efforts are the norm today. It's nice to have everyone in a single office for a few conveniences, but it's unnecessary for most stages of development.
There are too many good collaborative tools and methods and well-established processes that allow distributed project development, management, and project management. Use the best coders, and their location isn't really relevant, IMHO.
There is a both Darwinian and social pressure characterstic in the uncontrolled/unsupervised, sometimes playing ground of OSS.
Bounty is one thing; peer pressure another. In OSS, I've found that peer pressure counts. These are people that don't like to make mistakes (not that anyone does) and are more absolutist (yet creative). There is a gravity of attraction to the model that incents creativity while producing good code in both solo and collaborative environments/projects.
But as in most things, follow-the-money helps. This is why MySQL AB and other organizations with a business model can react deftly. That's not to say that apache.org is indecisive, rather that I put paying customers first ahead of charities. In the final analysis, FOSS is charity. And we all have to eat.
This doesn't excuse long bugs in a product, but it tells you where the product has been and is heading. Each app is different than the next. The common denominator, however, can be quite high, and visibility counts. So does cash.
Many apps point to the kernel, or a long-time, non-mutating set of core apps whose functionality and reaction are pretty well known in Linux and BSD. You can even look inside the code to see what's going on. The location of the coders is irrelevant.
What's relevent is inter-application dependencies, API sets, library cohesiveness, and the monolithic differences among apps. Inter-application communications happens at all sorts of layers, using differing sets of APIs and data structures in different states. These all have an effect on what must be done when a change is necessary.
It's not a house of cards, but rather like girders in a building. You can change one or two at a time, but not a lot of them without destabilizing the process or creating too many instances of change at a time. Change too many variables, and stability goes to hell.
There is a weakness in the open source process where rigorous and commonly/standardized processes aren't used to correct problems in code. While the CVS methodology is a nice archiving practice, what's missing is a standard process of reporting bugs-- and dealing with them in a uniform, rather than ad hoc method.
But making source visible means that the process of modification is more hive-like, rather than boundary-laden. So many can look and consider, rather than one business unit at a time, which slows down the process and creates multiple dependency instances. And, when you donate and fund coders, they seem to often move more quickly. Did you press the PayPal button?
Patches, no matter what they are, are woven into most things that Microsoft and developers do. There are numerous dependencies, and the numerous divisions, API sets, and partner dependencies make this difficult if even impossible to do on an ad hoc basis, as a generally available patch that breaks things is irresponsible.
Yes, it happens anyway.
Thie is the downside to having a huge, inter-dependent set of apps. Regression testing and dependency testing regimens have to be followed to ensure that small or even massive destabiliations don't happen. This also means that the easy stuff and the most urgent stuff (by their reckoning, not necessarily mine or yours) gets done first, and the tough stuff is just tough.
It's also what makes the closed source model more difficult to deal with, as Microsoft isn't just one pool of programmers, rather thousands of coders working on largely interdependent projects. While it looks like they should be able to do this, it's a reality that it cannot. And it would be irresponsible for them to do so, given so many users, and so many inter-related apps. We just wish it could. That's why OSS methodologies have a bit of an edge in this context (and others).
Historically, the book makes interesting points. There is another world outside there, where Microsoft has no part. Their stuff can be enticing. Still, there are platform issues, security issues, coding issues, just like every other OS.
But NO other OS vendor has been convicted across the world, or settled out of court so many times, than Microsoft. Their intent isn't benevolent code development, rather, profits. It's ok to make profits, but not to engage in the long list of bad behaviors they've been convicted of.
Is FOSS inherently better? In numerous ways, it has advantages, especially if you code in C and are schooled or have learned *nix derivatives. I have, but my brother has no chance whatsoever of learning *nix derivative culture and doesn't really care to. He just needs to make a few apps work. A few distros provide that opportunity handily. So does XP.
It's ok to take an objective view of the *nix vs Microsoft argument and choose accordingly. But such a sensible approach is usually eschewed here for fighting, brutality, code-righteousness, and anecdotal information. I hope his book does well.
These are people that told us that OS/2 would be king. Their veracity is in constant question; they make a living from making corporations believe that Gartner has information that they don't. And virtually all their research is vendor sponsored. Add this up: for Gartner to admit that Linux has such a huge PAID FOR share is tremendously good news.
Sales? Consider: when did YOU last BUY Linux? Use Fedora Core? Most versions of SuSE worth having? Most of us don't pay for Linux, although it's nice to have Red Hat or Novell, or heaven forbid Sun to fall back on.
If you RTFA, then understand the apparent pro-Microsoft bias on the part of the columnist at CTZ, you'll understand that there's actually very good news here: people actually paid, as in money, for Linux to get support-- because you don't have to buy it any other way. They did this, in competition with Microsoft's value proposition. That says more, and positive news about Linux than ever before.
Will it scare humanity into changing their habits? I would hope so, but the US ignores the Kyoto Treaty, and burns CO2/CO-producing fuels at hell-bent rates. Mass transportation? Nah.
It proves that unless you're interested in murdering subsequent generations, we need to start now to get energy that doesn't smut-up the atmosphere, our lungs, and forestry/ag plans that don't cut the lungs out of the earth so that someone can have cute cabinets in Miami.
Unfortunately, a little more natural drama (maybe a few dozen more hurricanes this year?) to get the body of humanity to change their habits.
Slashdot Mirror
Instead of creating meta-machines that serve a lot of users, home and otherwise, he's correctly pointed out that machine makers aren't serving other markets, like having cogent product line including mobiles, PDAs worth a crap, and so on. Computer makers don't understand the CE marketplace at all, and so it's no wonder they can't make money at it.
Michael Dell was the first to figure it out, but others have had moderate success driving consumer features. What they can't drive is consumer operating systems, because there ARE NONE.
Maybe one day....
isochonous applications are here to stay, including real-time video, lots of VoIP, and others whose time-domain is sensitive to latencies. The expertise of the writer involved is suspect, quoting Son-- who ran Softbank and isn't an engineer. Routing protocols, QoS, route saturations, re-authentication cycles, all of these cause objectionable latencies-- not to mention the end-2-end capacity of the network involved. Yes Martha, we need bandwidth. Ignore the idiot.
Case: email Case: IM Case: online gaming Case: forums Case: surfing Case: RSS Case: interactive purchasing Case: downloading entertainment Case: blogging Case: social infrastructure Go on. Run the sieve. Tell me what's not addictive. We're social and interactive creatures. Ask the question again. What knid of dumbass question is this? Yo: Cowboy Neal--> learn to ask a reasonable question.
So very cool, those dishes. Now: 802.11g at 45million miles. Go on, make my day. I want to crack WPA on Mars!!!!!
because that's one helluva Pringle's can. Defcon contests, you're over.
No load charity is a wonderful thing. The current model of huge organizations with powerful overheads taxes donations. If you consider the actions of the administrators of charity, they're much like a program in numerous ways. I'm not saying that eliminating administration can be perfected (you need audits, and program adjustments along the way), but a higher efficiency model would be welcomed. Bureaucracies tend to grow until they suck away the intent of charitable organizations. Some are far more efficient than others.... but in a future world perhaps more of contributions gets to the need, rather than the machine that services the need.
The whole Java hegenomy is a different culture for different apps than LAMP/SAMP etc. "So 90's...." doesn't talk well about platform compatibility, resource management, coding methodologies, and numerous other characteristics. It's like comparing an airbrush to an airchisel. Ok, bad metaphor but also a bad comparison.
Maybe the correct phrase is a hegenomy of devices, as this is what we have.
To extend your checking flights metaphor, I can do this on my mobile, my PDA, my notebook, or a terminal somewhere I don't own (not that I would). Each device is running something different. The mobile runs Symbian; the PDA runs WinCE, the notebook runs MacOS, and only heaven knows what the public terminal has, probably a Windows session.
At the core on the thick side is (statistically, anyway) either Apache/Tomcat, or IIS/something running the back end. Maybe Solaris, HP/UX, or something else is behind the curtain. Sun is trying to sell what's behind the curtain without thinking about the rest of the capability of the delivery system or the end device. Indeed the end device should go away or become something very uniform and manageable by their last perceived closed app, Java.
Yuck. I don't think that behind the curtain model works. Yes, hulking fast servers are good things. But divorcing what's at the edge is really very silly, unless you're a hardware server maker like Sun--- who provides none of those edge devices-- so in their minds they must not exist. These devices aren't embraced, they're ignored. It's egalitarianism through ignorance and hubris.
This is the same thick model they've been bandying about since inception, and failing-- except during the dot-bomb era when people just bought hardware for mindless reasons and irrational exuberance. As Robert Plant might sing, the song remains the same.... just a new stanza.
Sun is otherwise pretty smart, and smarter than Red Hat and SuSE when it comes to Unix. But they're also stuck in their own mud. McNealy and Schwartz should exit, and get a team that can appeal to a new and differently incented group of buyers. But their egos get in the way. They always do.
Casting TW as AT&T just doesn't work. Here's why:
1) AT&T and Western Electric were a monopoly, as in NO ONE ELSE GETS TO PLAY. CPE, as in phones, PBXs, and so on, were 100% controlled until PBXs broke loose, then the Green Decision to break them up. Their various groups were split into pieces that followed business unit logic: Lucent, AT&T Long Lines, and so on. History marks this
Time Warner, on the other hand, is a media empire with broadband holdings. In some areas, BrightHouse and other subsidiary divisions provide entertainment/TV Cable, and adjunct to that, broadband access. Their penetration is good in the US, and the US only. AOL was once the king of dial-up. No more. Dialup isn't a good situation. But as you can live without a phone, you can live without broadband access. Cable TV is like a utility, and uses utility metaphor, and has been promoted legally to the sense of a utility in a handful of markets.
But TW has serious competition in each business segment. Comcast, Adelphia, not to mention Verizon, SBC, BellSouth, and so on. AOL has competition: MSN, Google, Yahoo, and numerous others. TW publications have competition across the board. Their publishing unit has excellent and cogent competition across the board. THIS IS NOT A MONOPOLY.
2) I applaud the breakup of TW for numerous reasons. It will make Brighthouse stronger, and not beholden to the rotten cash problems that TW has. TW has demonstrated time and again that they're not leaders, and barely followers in this area. AOL needs fresh blood and new ideas. Their business model made sense when broadband build-out was small and spotty. It's not anymore. The added value needed for AOL would be to buy something like MySpace and add put it on the google revenue model. But AOL is leaden, and Steve Case is gone, and I think personally that they'll slowly lose share until they croak or are sold off for assets.
3) There are no other parallels that exist in the comparison. It's ok that to say, "gosh, bad idea". I have to do this, too.
If you want to have a discussion on the problems of national broadband inconsistencies, or how publishing content and revenue models now have become divorced from providers, or how IPTV will disrupt the living hell out of everything, then lead on. But it's an apples vs durian discussion here. They might be both businesses, but it really stops there. And like durian, AT&T stank.
4) Competition? Every telco. Every cable company. Every magazine/book publisher. Every content generator. There is no current problem with broadband competition. There may be as, like the recent 2600 Magazine cover implies, a new Bell, as they seem to re-unite again. But the telcos are only now putting in fiber, and worse, passive fiber. The last mile is still an open book, and the chapters are by no means written yet.
So, no. GIGO. SHOUTING IN CAPITALS DOESN'T INCREASE THE COMPETENCE OF A POST.
This goes further into the model proposed by the post: 8-cores in the CPU (one FPU for the eight) and 32 discrete threads, all in a 2U server box. This is based on UltraSparc, but there's Solaris 10, and the port of gcc to it for seductive app transfer. The whole idea is a hardware play.
It makes me wonder why there must always be this gulf between hardware and software vendors. The most successful models meld them together handsomely into devices like iPods, mobile/pda devices, etc. This thick-thin shift is so insane. At the end of the day, we just want to do work, entertainment, and something useful with the devices we buy, and the location of what's going on is increasingly irrelevant. But perhpas this is what (F)OSS software will get for us, an army of coders coupled to an army of blade vendors, with dumb devices at the edge.
Ma Bell, when she was an aggregate whole, had over $4B in assets in 1972 related to ROW, easements, and property that was a sanctioned utiility monopoly.
Time Warner has assets, that in 2005 dollars, are but a fraction. Indeed most of their distribution infrasturcture rides other utility ROW and easements. They are by no means a monopoly, although they do enjoy franchise status in numerous areas. That franchise can be revoked if the company is re-capitalized. Federal law also supports numerous re-examination possibilities when they recapitalize, or alter their structure, or are merged. It's non-trivial.
Equating it to the break of Ma Bell just doesn't stick to the wall. The issues are just not the same. I analyze the comm/telecom industry deeply. Regretfully for the poster, they aren't equivalent, not even slightly.
Comparisons to the Ma Bell breakup don't work here; that's the point.
Icahn is trying to up his asset. Ma Bell was a utility, and a monopoly. Time Warner is neither.
Time Warner has diverse media and manufacturing assets, and while they have content generation and distribution, they're not a monoply in any of the areas of their business, like Ma Bell was.
A breakup in this case, has no useful metaphor to Judge Greene's breakup of the AT&T components into different companies. The only common denominator between the two, is that the fact that they were broken up to increase shareholder value. Other comparisons aren't valid. This is an economic issue, not one with parallels.
Comparisons that try to link the two in that way just don't work-- TW's asset structure, business model, governmental regulatory control, and every other facet between TW and AT&T are different. Really different. The supposition of the post tries to presume that tie, where no tie exists.
There is no dependent public service on Time Warner. All they things they do, especially in media, can't be considered in the same way that the monopoly of AT&T and their FCC and PUC/state-governed telephony was.
These two aren't equal; they aren't congruent, they aren't even parallel.
Time Warner's broadband properties are not a utility, like water, sewer, electricity, or natural gas-- things you can't live without. They don't have the same history, the same economics, the same monopoly control, the same easement and right-of-way capital assets, and so on.
Therefore, garbage-in, garbage-out. The comparison is null, and it is, unfortunately moot.
There are people with spinal cord and brain injuries that could really benefit from this. People with trauma recovery with no sensation that don't realize an appendage is bumping into something. It's not quite like having skin with nerve endings, rather it's an early alert that senses something that you either can't, or can't *yet*.
In another way, it's also a way to help people recover from muscular atrophy, sensing leg movements, or arm movements. It can tell you when something's too tight, or incorrectly applied. Think physical therapy, or improving your golf swing, football kick, or reducing RSI.
Although I don't understand its resolution capability, it could also be used for carpet-fabric that could tell people when someone's at the door, or that someone has been in a room, or that the person weighs 100kg, etc.
Use your imagination beyond sex. I find this fairly fascinating.
There are already appliance makers that do this very thing: identify malware and viruses, and signal the others, usually in the guise of spam control appliances.
Webs of early notifiers is also not a new idea; look at the honeypot networks that are on the web, the honeypot project, and so on.
The containment cited is theoretical, subject to the ability to correctly identify behavior, and doesn't prevent users from clicking on URLs that have malware, or filter signatures that have fast breakout behavior.
And so, the merit of the Nature article is in question. It's just a PR release in disguise.
My observation is that you're a coder. Few others write such long sentences.
There are cogent arguments supporting micro- vs traditional kernel designs. Microkernels are valuable when you control hardware architectures, or where one goes heavily to assembler to make a CPU sing. They're less fragile, but also mandate are more serious hardware examination approach to processor optimizations. That's why Apple can do it but it's tougher for others. Their jump to Intel proves that the Darwin kernel approach first perfected in Mach is actually not as tough to remake onto other platforms. I've been speculating that the OpenDarwin tree was more than just a curiosity on Intel/AMD platforms. It's the Trusted Computing area that pushed the decision over the edge, in my belief. Jobs has foud the mirth of making Hollywood happy. Jobs moved the Carnegie Mellon work done on Mach to Apple to be manifested under the Apple design methodology, and did so successfully after a few mild, non-showstopping mods. The result has been positive. Adding iTunes was astute.... four times he's done this, first with the Apple ][, the Mac, the MacOS, and now the *Pod designs. But it isn't for all, and simply points to the successes of an avatar.
Solaris/UltraSparc could take advantage of the same model... but they also own keys to Unix and BSD-- hard fought and won and politically continue the SunOS legacies in Solaris/Open Solaris.
Longhorn will be interesting to see. At some point, Microsoft must balance legacy architectural mistakes made in the fast-and-loose era and hone a better core architecture. This will be tough for them. But it'll be less expensive than the madness they face now. Their security models have improved vastly, and the raw brainpower purchased with Connectix, ZoomIt and others is paying handsomely. But it's closed, and bereft of numerous allegiances that power (F)OSS. The concept of free rocks them, and Google's kinda-free methods are infecting them. It'll be interesting to see what happens in five years.
A slight historical background is in order.
The DOS->Win31->win95-win98-winME family is ideologically and in reality built on DOS. This model wasn't ever a good one for multi-user anything, and was constrained to hybrids of the peer model rather than LAN/WAN hierarchical resource utilization/RPC models.
The OS/2->Winnt31/35->NT4->NT5/Win2K/XP/Win2003 is a different family and was better, hobbled by a legacy addiction to the IBM legacy models of OS/2 and the PDC/BDC/LAN Manager legacy. Windows 2000 evolved the Active Directory concept that put still another nail in Novell's coffin (please don't go to NetWare vs AD, etc.).
The end of the story says that you're right, but partly for different reasons than you state, IMHO. Microsoft did evolve object classes and API sets that are pretty clever. But when so many are strung together, and become dependent, it causes enormous delta in what can be easily changed and evolved. That's why new products and releases are so leaden in getting through the dev process. This is also coupled to the fact that other developer partners are loathe to spend money to constantly update code-- and this is the motivation behind Microsoft's stated goal of major releases every four years with revised kernels within the same release every two years. Look for "R2" of Win2003 server coming VERY soon.
Linux, by contrast, is done when its done. It views collective substrate modifications at different levels.It also evolves older editions of the kernel (e.g. the 2.2, 2.4 tree development) as active projects and maintenance obligations.
The Darwin kernel that underlays MacOS is different still, but is a microkernel approach and requires even less dependency checks because once a port is done, the compiled kernel fixes rarely have an effect on core API sets and distributed (mostly BSD) applications. Apple controls its developer network much better, and developers mix BSD, GNU, and other OSS and OSS-like models in a blend/blur.... some open some closed. But Apple carefully controls the microkernel, and so far, microkernel patches in the past four years have been surprisingly few, and Apple's problems equally few from a security/functionality patch perspective.
Solaris has similar qualiities, and now that gcc works well on Solaris, ports of various apps have found a new home on a stable environment.
The open design of Open Solaris, many BSD branches, and Linux have various benefits here. And, as seen in another reply, occasional lapses based on simple motivation to fix bugs.
Collaboration is locationally independent. Lots of on-line whiteboards exist, if and where needed. The lingua franca is English, and C, maybe with some perl and php thrown in.
I like EU coders, Balkan coders, Indo-Pakastani coders, Chinese coders, Aussies, all of them everywhere. Coding is a pretty personal exercise, even in teams. Collaborative efforts are the norm today. It's nice to have everyone in a single office for a few conveniences, but it's unnecessary for most stages of development.
There are too many good collaborative tools and methods and well-established processes that allow distributed project development, management, and project management. Use the best coders, and their location isn't really relevant, IMHO.
There is a both Darwinian and social pressure characterstic in the uncontrolled/unsupervised, sometimes playing ground of OSS.
Bounty is one thing; peer pressure another. In OSS, I've found that peer pressure counts. These are people that don't like to make mistakes (not that anyone does) and are more absolutist (yet creative). There is a gravity of attraction to the model that incents creativity while producing good code in both solo and collaborative environments/projects.
But as in most things, follow-the-money helps. This is why MySQL AB and other organizations with a business model can react deftly. That's not to say that apache.org is indecisive, rather that I put paying customers first ahead of charities. In the final analysis, FOSS is charity. And we all have to eat.
This doesn't excuse long bugs in a product, but it tells you where the product has been and is heading. Each app is different than the next. The common denominator, however, can be quite high, and visibility counts. So does cash.
Many apps point to the kernel, or a long-time, non-mutating set of core apps whose functionality and reaction are pretty well known in Linux and BSD. You can even look inside the code to see what's going on. The location of the coders is irrelevant.
What's relevent is inter-application dependencies, API sets, library cohesiveness, and the monolithic differences among apps. Inter-application communications happens at all sorts of layers, using differing sets of APIs and data structures in different states. These all have an effect on what must be done when a change is necessary.
It's not a house of cards, but rather like girders in a building. You can change one or two at a time, but not a lot of them without destabilizing the process or creating too many instances of change at a time. Change too many variables, and stability goes to hell.
Find me a bug-free piece of sophisticated code.
I dare you.
There is a weakness in the open source process where rigorous and commonly/standardized processes aren't used to correct problems in code. While the CVS methodology is a nice archiving practice, what's missing is a standard process of reporting bugs-- and dealing with them in a uniform, rather than ad hoc method.
But making source visible means that the process of modification is more hive-like, rather than boundary-laden. So many can look and consider, rather than one business unit at a time, which slows down the process and creates multiple dependency instances. And, when you donate and fund coders, they seem to often move more quickly. Did you press the PayPal button?
Patches, no matter what they are, are woven into most things that Microsoft and developers do. There are numerous dependencies, and the numerous divisions, API sets, and partner dependencies make this difficult if even impossible to do on an ad hoc basis, as a generally available patch that breaks things is irresponsible.
Yes, it happens anyway.
Thie is the downside to having a huge, inter-dependent set of apps. Regression testing and dependency testing regimens have to be followed to ensure that small or even massive destabiliations don't happen. This also means that the easy stuff and the most urgent stuff (by their reckoning, not necessarily mine or yours) gets done first, and the tough stuff is just tough.
It's also what makes the closed source model more difficult to deal with, as Microsoft isn't just one pool of programmers, rather thousands of coders working on largely interdependent projects. While it looks like they should be able to do this, it's a reality that it cannot. And it would be irresponsible for them to do so, given so many users, and so many inter-related apps. We just wish it could. That's why OSS methodologies have a bit of an edge in this context (and others).
Historically, the book makes interesting points. There is another world outside there, where Microsoft has no part. Their stuff can be enticing. Still, there are platform issues, security issues, coding issues, just like every other OS.
But NO other OS vendor has been convicted across the world, or settled out of court so many times, than Microsoft. Their intent isn't benevolent code development, rather, profits. It's ok to make profits, but not to engage in the long list of bad behaviors they've been convicted of.
Is FOSS inherently better? In numerous ways, it has advantages, especially if you code in C and are schooled or have learned *nix derivatives. I have, but my brother has no chance whatsoever of learning *nix derivative culture and doesn't really care to. He just needs to make a few apps work. A few distros provide that opportunity handily. So does XP.
It's ok to take an objective view of the *nix vs Microsoft argument and choose accordingly. But such a sensible approach is usually eschewed here for fighting, brutality, code-righteousness, and anecdotal information. I hope his book does well.
These are people that told us that OS/2 would be king. Their veracity is in constant question; they make a living from making corporations believe that Gartner has information that they don't. And virtually all their research is vendor sponsored. Add this up: for Gartner to admit that Linux has such a huge PAID FOR share is tremendously good news.
Sales? Consider: when did YOU last BUY Linux? Use Fedora Core? Most versions of SuSE worth having? Most of us don't pay for Linux, although it's nice to have Red Hat or Novell, or heaven forbid Sun to fall back on.
If you RTFA, then understand the apparent pro-Microsoft bias on the part of the columnist at CTZ, you'll understand that there's actually very good news here: people actually paid, as in money, for Linux to get support-- because you don't have to buy it any other way. They did this, in competition with Microsoft's value proposition. That says more, and positive news about Linux than ever before.
Will it scare humanity into changing their habits? I would hope so, but the US ignores the Kyoto Treaty, and burns CO2/CO-producing fuels at hell-bent rates. Mass transportation? Nah.
It proves that unless you're interested in murdering subsequent generations, we need to start now to get energy that doesn't smut-up the atmosphere, our lungs, and forestry/ag plans that don't cut the lungs out of the earth so that someone can have cute cabinets in Miami.
Unfortunately, a little more natural drama (maybe a few dozen more hurricanes this year?) to get the body of humanity to change their habits.
But we can hope.