After developing applications for a wide variety of banking industries it became clear that:
1) The only way to develope software systems, is to proactively secure the systems once they are deployed.
2) To proactively and continuously review and examine such systems, you must have the source code and build tools and access to the hardware engineering requirements of the systems involved.
3) The only known process where this can be achieved is through Open Source.
Closed binary proprietary software is not secure, cannot be MADE secure, is impossible TO secure and with patents and copyrights laws as written it could be quite possible you could be SUED for securing the software yourself.
Security became an extension of the software engineering process for the company I started previously, and it involved reviewing the source code and making changes, performing attacks, etc.
Critical to this process was to have as many eyes and opnions looking at the source code as possible. The more experienced professionals that had a chance to offer advice and opinions on the code, the better and more secure the code became.
An entire portion of the software engineering process cannot even be done with proprietary software, and I personally as a CIO, declared proprietary binary only software sales DOA in this industry 2 years ago.
Primarily, if you are using the "experimental" options in a kernel build.
If you attempt to add experimental options in a kernel that uses a different kernel config file, the results are undefined.
That is, the kernel isn't guaranteed to build correctly with experimental drivers, adding a diffeent config file with experimental options or built in options that need to be configured at compile time, could result in values that are compiled for modules that could cause an oops.
A good example, is the tag settings for a experimental SCSI driver. Usually you can set how many SCSI commands you can have outstanding, for your SCSI bus, before the device demands attention again from the driver.
Say the author put in a value of -1 for the experimental driver in release 2.4.x.
You makde a config for that and put it in a recent release, called 2.4.x+1.
Well, 2.4.x+1 is no longer an experimental release, and -1 when compiled, will be outside the range of the expected values. The Author, expected you to generate a new config, which WOULD have provided the module with the proper info.
So you compile it, and when the final release driver gets a -1, it oops the kernel, trashing your partition.
You know, I talk about the state of software with a large number of companies.
All of them clueless, specifically, about what is about to happen to the IT industry/software industry world wide due to the open source movement.
The biggest losers? US and Europe?
Why is that? Continue reading to find out why China is the next 21st century super power, and the US and Europe will be side players for the most part in the later half of this century.
At the moment I give open office presentations, if I can get an audience, to companies and hope to have a multimedia version of it built so you can download it at my website: http://www.aesgi.com
But, here are some of the things I point out:
1) The computing industry, historically required companies and corporations of individuals, so that capital could be organized along with labor to produce software.
This was due to the fact it was incredibly expensive to obtain hardware, and software was a real bitch to build.
You HAD to have a corporate entity, or government backing to build software.
2) As our industry has evolved, hardware has become extremely cheap and powerful...through competition. Software Engineering has allowed tools to progress that enable extremly sophisticated views of the software process to become possible.
This is because hardware has become SO powerful, that and individual now can now rely on whole "environments" to make very smart decisions in constructing software, or at the very least, inquire about contributions to said software from a colleague.
3) The internet has eliminated the need for corporations/institutions to build high quality/very complex software. For very little cost, organizational units that are far superior to what can be bought can be assembled at very little cost using the internet.
Intellectual capital using this method of organization and procurement is superior to what any single corporation can purchase.
4) The corporate mantra/dogma of build software, release software for sale, then add features, then sell, then add features then sell again, is dieing.
Users don't want more features for often than not, when they only use 10% of a program's feature set anyway.
If all you do is write software, and all your revenues come from only sale of that software your business will start to die very soon.
Whole countries are rebelling, due to the fact that software costs are increasing, not because they have to, but because of Western political interests which seek to lock the market up with unfair laws, and to maintain the status quo.
5) Intellectual Property and Copyright law is about to undergo a MAJOR shift.
American companies are enacting Digital Copyright protection mechanisms that are fundamentally unsound to free trade.
Foreign countries outside the western sphere of influence know this, and are drawing up plans to protect their interests.
6) Countries around the world are watching the US and Europe. They are taking steps to prevent western influence of technology in their countries to protect their soverienty.
Specifically, using open source software, so that American political interests, cannot shut down a countries computing/information infrastructure should someone declare a software or patent violation of foreign soil.
7) China and its neighbors, have told the US and its allies that current copyright law and patent law as written is not acceptable to Chinese state interests.
China has taken steps to halt and reverse the progression of American/European Software into their country.
8) The Far East, the worlds most populous nations, are building infrastructure, based on Open Source.
This has a number of very important implications, such as costing American/European companies 1000% more than Asian companies, just to route a packet over a network, or do a simple mail merge with your favorite office suite, to simply storing a file or printing it....the rest of the presentation, I hope to have up for download soon. I am working out the kinks with the audio make overs for the slide show, but hope to have it all done soon.
China has made a few things clear with respect to its rise as a new super power:
1) We will show the world what can be done and take the lead as a nation in engineering, space exploration, and computer science.
They started with the largest engineering project ever concieved..the largest hydro electric power project EVER.
No western nation could duplicate such a project even if we wanted to because of the sheer size, and the use of cheap labor by the Chinese.
China will have clean, cheap energy for 25% of the nations total needs for the next 100 years from this project.
2) It has been rumored, recently in a M5 visit to China, that space exploration plans include a moon base within 15 years of thier first successful manned orbital launch.
It was also made in not so many words, that once they complete this base, we are NOT welcome.
3) China, doesn't want Western IT technology. Especially Windows, or Intel's chip technology in any sort of influence on its internal consumer markets.
It was made clear that Intel's Chip ID technology and the CIA's insistence that back doors be placed into foreign copies of Windows, was not acceptable.
It is forbidden in China to attach any Intel processor based system to the internet with Chip ID technology of anykind.
This new processor was a "call your bluff".
After all, HOW DARE YOU compete with Western technology, you can't possible build anything near as well as we can, so you must accept our processors if you want to do business with us.
It is this direct response to Intel's Digital Rights, chip ID technology as well, that this processor now has been born.
It won't take China long to ramp this processor up to Pentium 4 Xeon quality (3 years at most), using Tainwanese acquired fabrication planets and technologists.
In my view, China can sustain its economic growth internally, due to its population size for the next 50 years, and tell the rest of the world to kiss its ass.
I believe greed, our very own Intellectual Property Rights Laws, and this obsession with Digital Rights Management has locked us out of the only market that will provide long term economic security to the United States.
What is more, I believe China is using these laws against our very own interests in doing business over there.
Coupled with thier own version of RedFlag Linux, once China has its own PC, and own Operating System, THEY WILL DICTATE TERMS 50 years from now to the Western world.
Our business leaders are fools in this country, and while China builds the biggest terrestrial project ever concieved, builds its own Moon Base 20 years from now, the world will do everything it can not to make the Dragon DISPLEASED.
1) Trying to build a not so average publishing API, with some truly remarkable features in Java, using parallel databases.
http://www.aesgi.com
It works, but it has issues, and what is more, its fast.
2) Talking to companies about Linux, on average I spend about 20 hours a week at corporations discussing Linux...
I have to digest enourmous amounts of data to build these presentations, so the amount of research I do approaches about 20 hours a week.
If your in the Madison area, you ARE on my hit list.
3) I run 6 miles a week, or about 5 hours a week doing something healthy for myself.
4) I sleep 8 hours. Very important.
Plus I eat like a pig, but thats OK because I am blessed with the metablism of a small gerbil or equivalent rodent. 5% body fat! HA HA.
I am the disdain of women everywhere when I go out to eat.:-)
5) I am building broadcom chipset drivers for my 802.11G card for my Linksys access point.
Surprise, you can't get braodcom drivers for Linux.
Probably because Microsft has paid them under penalty of IP to make them specifically for Windows...
ONLY.
But I am reverse engineering them using a windows box.
HA!
Just try and stop my from releasing them soon Billy Boy!!!
6) I have to write proposals for deploying solutions I come up with, thats another 8 hours at least.
7) I am a LLC in the state of WIsconsin so I have to do quarterly tax reports, plus handle 1099's plus, I am considering sponsoring a H1B Visa guy, who is a good friend of mine, so he can become a US citizen. Got to do all that business stuff too...accounting, setting up my own computers, etc.
So, I am CONSTANTLY working, and if I may say so I get it all done on time under budget (mostly:0) by:
Buy a laptop. I can work anywhere and do.
The ole office/apartment can get opressive, so I go out and work in restraunts, the student UNION, Barns and Nobles or Nicks Bar on state street....etc.
Since I have a sager notebook. I get about 40 minutes of battery life, so I have to pick places I can steal electricty from.
Point is, I work constantly, no rest for the weary...or in my case....the wicked.;-)
I am surprised nobody wrote me back on how the SOFTWARE INDUSTRY is thinking the same way as the Biotech industry is in the US.
(i.e. You CAN'T OWN your own software, AND you MUST RENT software from us or it is illegal.)
That way, you never write a piece of software that is of any kind of high quality. This enables you to sell defective software, and continuously charge your customers for defect corrections.
Which, is a relatively new concept, thanks to Microsoft.
If we didn't have so many lawyers involved in the Biotech community, we would HAVE growable organs NOW, and very much less ethical considerations in who to hack up and why.
Besides, you may be able to live with a donated organ, but donor patients are soon wishing they were dead with the medical bills which they will have to pay for the rest of thier lives, the enourmous health problems associated with depressed immune systems, and the Frankenstein like chemical cocktails they pump into your body to prevent rejection. Many of these chemicals, are poorly understood how and why they work.
Besides, which brings up another point I frequently make. Why would any private medical company EVER want to cure ANYTHING?
Think about it, it IS NOT profitable to cure people. I mean, if you cold grow someone a liver from their own genes they are cured!
What is the profit in that? I would concentrate on doing better transplants, so I could sell people all kinds of drugs, and tons of extra medical care for a transplanted liver.
Not only that, they have to keep buying the drugs from you for the rest of thier lives!!
If they can't pay they die!
Profit or else, or else if you refuse to pay, you die. What a concept!
I think I will patent the idea.:-)
But since only 1 or 2 companies can work in in this area of growing organs, due to patents, it is going to take a very long time indeed before we see any organs you can grow.
(i.e. Most of the handful working working on this problem are idiots, and pretty slow learners as well, I am afraid.)
Besides, like I said, there is now way a board of directors is going to allow growing organs from ones own tissues....it doesn't make good business sense...for the masses anyway.
Now, if your that very rich person on that board, well, of course you can grow your own organs.
I use to work in Biotech and I can tell you the stuff that is considered in closed rooms, shut from public eyes, would make the slashdot audience roll thier eyes in disbelief.
Those being questionale in my opinion with regards to BK and the Linux Kernel, one based on GPL code and the other not.
Long term these two cannot exist. Why?
1) The primary reason why is that companies in what I call the old "established" economies of software manufacturing, insist they have complete control of the software.
2) The second reason is due to the fact that as we transition, and leave these companies behind, to the new economic reality of "software commoditization", said companies will enact/engage DMCA/Copyright and IP laws to prevent competition or improvements to the process of building the Linux kernel.
1 is obvious. 2, not so obvious. I mean, what would happen if for some reason the Linux kernel couldn't be developed, or not easily developed, on something besides BK?
Furthermore, BK isn't the end all be all of software development methodologies to handle source.
You have to remember, that BK and the company that owns it is basically TELLING YOU this is how we will develop and manage software.
Which brings up my final point:
3) All software in the proprietary space is being usurped gradually by engineers who are building better software, through a process that continues to evolve.
This process isn't tide to a particular piece of software that cannot easily be changed, to keep innovation of the basic practice of building said software from stopping due to a license restriction.
I think I have to disagree with Linux on my final point. The Linux kernel will soon reach an epiphany, probably around Version 3.0. Which will require some fairly advanced pieces of tools to be developed to keep development, and the community on track.
When we reach that place, we will need tools we can modify to bring the process of building the software, as well as the kernel technology itself, to new levels.
I don't see how that is going to be possible, if we continue to run BK.
If this continues, kernel timelines, and code quality will suffer. Already 2.6 is an immense challenge.
We should be looking at, quite perhaps, something like and Eclipse style plugin, or perhaps a seperate tool for KDevelope, to handle specific kernel versioning, and development issues in a standardized environment.
Right now you have to spend a good 30-40 days of continuous reading, kernel track, source code, and putsing around with building perhaps 1-2 machines to build anything serious for the Linux Kernel.
After 2.6, we will need some sort of IDE to get programmers up to speed faster, so they can learn more quickly, about the kernels innards, etc.
We don't need to rebuild an IDE, specific for kernel programming, we could design Plugins for Eclipse or KDevelope to further that task.
What could be included would be:
1) A current and relevant deprecated code completion aware editor. Perhaps, connected with a version control system that would keep developers aware of changes in the main development trunks.
2) A more complete robotic/remote host plugin for developing/stepping through those nasty sorts of device drivers you pretty much need two machines for: 1) Video Drivers and basically Network drivers, for an adequate building environment.
I mean, given the responses from the BK owners, imagine if Sun told the Eclipse team: Don't you dare build a remote debugger, we will reengineer the protocols the the JDA (Java Debugging Architecture) with ever jdk release so you can't debug code without our software.
You can bet Java code would suck even more than it does so today without an integrated debugging environment.:-)
--
In conclusion, I think these sorts of things will be very difficult to create or build, if the very software that is holding the main development trunks of the kernel, cannot openly participate in the building of a continuously better approach, to creating/debugging/versioning/managing source code for the Linux Kern
There is nothing open or standard about Microsoft's.Net platform.
The entire platform wasn't even created to solve a business problem, it was simply created to kill Java.
Well, that and because of.Net's distributed DCOM/DCOM wrappers (which by the way that is all.net is, is repackaged COM/DCOM via the web) was invented so that you buy MORE Microsoft software.
Java was created to solve an industry wide, 30 year problem. How do you turn software from an expense, into an asset?
You do that by making the software independant of hardware, and making it a part of your IP property of your business process.
Which should be perpetual, respond only to a business need, and not a vendors whim.
The reason why Java was invented was to free companies from buying a PeopleSoft implementation for 10 million and then having your investment totally invalidated by someone who buys the company and destroys your investment.
Like Oracle plans to do.
Open Sourcing Java would simply free licensed implementations from being engaged by the certification process.
I am not so sure, that is great for Java. Sun has been extremely magnanimous in managing Java.
Having a bunch of Open Source implementations running around with no means of making sure they all work the same way is not what I would call a positive thing for Java...GPL'ed or otherwise.
Remember, Java write once run everywhere is critical. Perhaps not in the US, but too all my European and Chinese and Japenese customers, who are quickly moving away from anything Microsoft, and everything Linux, it is very very important.
Some sort of release mechnism will have to be developed to insure that Java GPL'ed implementations are not allowed to fork with non standard runtime libs.
since you are talking about FUTURE, and not now, I would first and foremost use Linux to manage the bandwidth. I assume you haven't made any buying decisions yet?
You can buy propritary heavy lifting equipment such as Cisco gear, and collapse the entire management tier into one switch, but I wouldn't.
It will work fine, BUT you can't reuse Cisco gear as desktop PC's, or even donate them to a school when they get too old.
My way of building things, you can reuse the equipment when its all done.
With that said, I would get a rack of i860 Motherboards, such as SuperMicro's with 6-8 Gigabit Ethernet cards in them and build a BGP based solution.
You get:
1) PC Hardware, commodity items. Easy to replace, easy to repair should something go wrong.
In fact, for some customers I have designed such solutions for, I usually put a spare unit in each rack, on the floor, and just shove it in when I need it.
In any case (4U), each rack would have about 10 machines, with one serving as a switching backbone. I would connect them all with Gigabit Fiber cards into one box. This box forms the QOS SLAVE. We had 6 racks of these, so I had 6 QOS slaves, that then were connected to a Backbone machine that monitored each racks incomming T3 Line, and rerouted the traffic from one rack to another based on how much was comming into each slaves backplane. Each of the 6 slaves was connected to the master, which had Nagios, and the master BGP/ASN lists.
The application was a video feed, from 8 different racks inside the same building, going out AND comming in from various sites. Voice over IP was also used over the same T3 lines.
Insuring enough bandwidth and prioritizing the traffic was key.
Most i860 PCI-Xpress (64Bit PCI) i860 Motherboards, especially the ones from www.supermicro.com can handle 6-8 cards in a chassis. Some manufactures sell external PCI-Xpress chassis and can accomodate up to 16 cards externally.
2) Linux Kernel 2.4.21 coupled with BGP, QOS additions to the kernel and a decent management program like Nagios, will give you all the bells and whistles you get with the big boys.
Except you don't pay millions for such a setup.
3) With such a system, you shold be able to build QOS ratios between BGP ASN's where you can alternatively route to a given pipe through a different rack that has the LEAST amount of activity on it. (i.e. in my case, that was, the least amount of T3 activity, easily monitored of course with Nagios and some simple SNMP scripts which updated the QOS tables, and if required BGP ASN lists on the master machine for the rack.)
You can adaptively route based on QOS signals and change your BGP asn lists to reflect traffic levels in almost realtime.
Been there, done that.
Ironically, my biggest problem with all this, was heat and backup power. 400-500Watt power supplies in a colo, that was originally designed for 4 U chassis with 200-300 watt power supplies, made the place pretty toasty. (78-80 degrees).
It is amazing how much heat the new Xeon processors and motherboards throw off, especially if you MAX OUT all the PCI-Xpress slots in them!
New air conditioning system, and an extra 5K main had to be shunted to the building. Total cost was about 32 Thousand. About 1/4 the cost of the entire project so that was a bummer.
After that the average tempreture dropped to 60 degrees inside!
I loved it, I could watch my pretty Nagios 3D graphs and eat my Diary Queen (one accross the street from the colo) Peanut Buster Parfie while watching billions of bytes go by, at least when I the customer called me to do work at the site.:-)
Undiscplined thought is a problem with many people.
If your thoughts tend to race, you need more time to slow down. Do you take the time during the day to focus or practice on making only one thought in your mind?
Like learning new things, your brain can rewire itself to learn how to think in this manner, if it is too "wiley", but it takes time and practice.
Mental concentration, or drugs that alter your neural chemistry, can also affect your intelligence as well.
You may be trading the serene mind, for a duller one, restricted by drugs, with fewer thoughts for one that perhaps with proper practice could be trained to control and utlize all those distractions.
Your body is a machine, and for the most part you have greater control over it than you can possibly imagine.
I would try far eastern techniques for focusing your mind for an hour or two a day for 6 months. If this doesn't help you, then try the drugs.
I think, several challenges to just materials science has to be overcome.
For example, our current science in engineering, relies on models and previous engineering attempts, to build new structures.
If you want to build a structure, say taller than the sears towers for example, you can do so, by using the Sears Towers as a reference, then building perhaps 10-15% taller.
Historically, we buld a large number of structures, not just buildings, a little bigger at a time. We build planes, a little faster at a time.
That is how our engineering science works. Even when we sent men to the moon, as colossal a task as that was, we took very small steps at a time, and it took decades.
Building something like a Space elevator, in the timeframe (10-20 years) I think is ridiculous given our current engineering science and application of Mathematics/Statics etc.
Just because you have a material than can go hundreds of miles straight up doesn't mean your structure will.
Whole new branches of engineering will have to be invented, as well as new mathematics to make this structure work.
Personally I think the work Stephen Wolfram has done so far in FSM's (Finite State Machines), may offer a clue as to how we can take much bigger steps in the sciences, with much more predictability, in our models, and methods of construction, to make a space Elevator possible.
At the very least his work sheds light on the principles of complexity, and why we take baby steps in everything we do.
Specifically, how can we design systems, when we have no working model, and to build such a model requires an order of magnitude in scale our engineering science, historically, has never had to deal with.
I think, after a century or more of using this material in terrestrial structures, to understand how it works better, we can start thinking about such an elevator system.
But I think it is a safe bet you are not going to live to see one anytime soon, much to the contrary some of these guys at the Space Elevator web site will have you believe.
My take on it, and I don't pretend to know both sides:
Some guy wants to be a partner after being at a company, for only 8 months, and only has what 6 customers?
I would say the guy is out of his mind, has no business sense, is unreasonable, or has an ego the size of Bill Gates Estate.
Let me tell you, that is one big ego.
Partner in the classic sense, means you bring something to the table nobody else has contributed to the company, that adds a great deal of economic value to the company.
That normally means something you can't get from hiring an employee.
If he had like 100 customers, or invested a million into the business, perhaps.
Being a partner isn't all what it is cracked up to be either...as the saying goes, becareful what you wish for.
For me, when I was upset with the last place I worked, working 80 hours a week, I left to be my own boss.
Now I get to work 80 hours a week, except I get to yell at myself for stupid things I do.
No wonder half the people I work with think I am nuts all the time, talking to myself.
That boss of mine is not only a whip cracker, he has high standards too!:-)
Something I sent to Slashdot Editors a few days ago...
Didn't get any bites though.
I think it would be a good idea, to bring a class action lawsuit against SCO.
I think we should use slashdot as a place to organize such a lawsuit.
For the following reasons:
1) I think technically, this lawsuit given the recent changes in management at SCO, involves fraud. That is, the companies officers know privately they do not have a case against Linux, and are fraudulently misleading thier investors/shareholders too personally enrich themselves with regards to stock price using a lawsuit to falsify product value to said shareholders/stock holders.
This is due to the recent in jump in SCO's stock price. The company simply isn't worth the current stock price, historically and is therefore artificially inflated.
2) Technically, I think, from the perspective of most Linux Kernel developers, including myself and SCO's own development group, that SCO may have abridged GNU code illegally. If this wasn't the case, I don't believe SCO would have continued to sell thier own distro after they committed the lawsuit.
The suit should include full disclosure of all SCO source code. Furthermore, seperate suits should be filed against SCO should GNU software be found in thier kernel.
3) The suit is affecting the industry, consulting firms, companies in real, economically negatively, in a measurable way. Customers are being lost, companies are having to spend money to switch, or consult legal people. This is all because of SCO's suit.
We do not need to wait to the end of this suit, we can file class action suit immediately to get damages/satisfaction.
I also believe that if we ajoin the company officers in #1, we should be able to file a seperate lawsuit against each officer of the company, and not just the company as a whole.
I think, we should use slashdot as a place to:
1) Ask people to generate documentation. Documentation of an official nature, which supports points 1-3. For example if you are a consultant, and you lost a job based on SCO's injunction and public statements, ask the customer to write a letter detailing the loss of business because SCO makes Linux too risky.
2) Internally, if you are working for a company, obtain permission to use corporate Email disclosures for any migration plans away from Linux.
3) Detail any personal damages as a result of not being able to make medical insurance payments, bankruptcies, or personal financial hardship as a result of the loss of business as a direct result of SCO's pending lawsuit and its affect on your salary/business.
I would be happy to help organize my time with regards to this, and would encourage anyone to Email from Slashdot as to how to proceed.
Please post.
Afterward, we can begin the process of selection of legal counsel once we organize.
-Gregory Carter -CEO -Applied Engineering Software Group -gcarter@aesgi.com
The IP laws fundamentally work against Open Source.
Any company can extract code from an Open Source project, such as Microsoft and then incorporate it into its product.
As such, IP law protects the company from this sort of illegal appropriation because of disclosure rules governing IP law and the DMCA act.
What we need, is something akin to the BSA and SPA. A "tattel-tale" website.
SPA encourages employees to tell on thier companys if they are pirating binaries.
Why don't we have such a website that allows employees to tell on companies that pirate GNU Source Code by incorporating it into thier products, and not contributing the changes back to the community?
After all, do to the enourmous amount of corporate corruption in the US, under the table political manuevers our #1 enemy is doing, there must be a huge number of burned out pissed off Microsoft/ACME employees out there.
We only need one.
Rewards would be part of the legal settlement, should money be awarded.
I think the primary problem is that the technology to build and design probes changes too quickly, and affects design.
I always thought that there should be a way, to build a probes navigation and propulsion systems in a standardized whay so that avionics software wouldn't need to change that much.
Sort of a standardized platform if you will for doing solar system exploration.
This platform would consist of a number of parts that would not change, and could be reusable in a number of different configurations for building a probe, depending on what its job was.
Cameras, photometers, spectrometers, and power sources could all be packaged in the same why depending on the probes job.
Every probe that nasa launches is always customized and built around cost and included packages.
I am not so sure that is the best way to go about it as you have to reinvent all the software to manage the probe every time you build one.
Probes should be cheap, produced in high volume, (thousands) and interchangeable.
With a standardized approach, failure rates should come down a bit and costs should be reduced.
I find a computer science degree exceptionally useful, for the following reasons, IF they meet the criteria:
You attend a University that contains the equipment and staff that will provide you with things you can't get in the private sector, that interest you.
This could be a wide variety of things...for me it was cutting edge research into the areas of vector processing/parallel compiler design.
To that extent I attended UW-Madison, which recieved a grant for 10 Million, in 1991 for a Thinking Machines 128 Processor machine.
I realized that there, I could get access to equipment and more importantly, Doctor James Larus, who had a large amount of work that he published on the topic that I could gain access too. I had a field day for 5 years.
But like you, left to start a company. Now I am back at the UW, 12 years later, finishing what I started.
So, my view on a Computer Science degree, and why you pay for it, is because you gain access to people and pieces of technology that you can't get in the private sector.
That is why I went to UW-Madison, that is why I returned. However, this is not always the case. Most colleges have very little in the areas of cutting edge equipment and "second rate" PhD's, that don't publish much or are not the leaders in thier fields. They are what I call teaching PhD's.
There are TONS of Universities that fit that description. They are still good, but just not AS good. Try and avoid these, but if you can't get into a good program, make the best of it.
Universities, also provide instruction into "foundations in science". What I mean by this is, although the tech you might learn might go out of style, Calculas never will go out of style. You learn calculas, deductive reasoning, and the rigours of the scientific method and research.
In the end, learning how to teach yourself very difficult abstractions and science, analytical thinking, will provide you with the means to avoid what I call "tech scams". The biggest one, is so called certifications in product families companies charge people for being paper experts.
These items Universities teach, don't go "obsolete". Calculas will be taught with the same fundamentals 500 years from now, for example...
Now more on the scams, once you get out of college....
However, I hold just the opposite view on Certifications. (CNE, MCSE, JCP...etc)
Certifications, don't provide any value whatsoever in my opinion. All they teach is extremely short term knowledge, about usually product families with lifetimes, say no more than 6 months or less on the market.
The Cisco, Microsoft, and related Java technology certification programs amount to memorizing product features.
I don't need to pay 10K and sit in a room with an instructor to figure out technology products. I use my own brain, Barnes and Noble's and some independant study after purchasing the product to become knowledgeable.
If they only way you can learn.Net quickly enough or Java, is through certification programs, I feel sorry for you.
Much better, in my opinion than some of the instructors who teach the class, I am actually using the product or technology to solve the problem, myself.
I can usually do it FAR FASTER, than purchasing a certification program too. I also feel much more confident. As well, I am not stuck someplace in a motel room for a week away from work either.
So, finally, with my requirements, I would say you are not getting a good deal, minimally, your basically getting a certification level type of knowledge just for a piece of paper.
If you really wanted to though, doesn't sound like your getting much, so I hope your not paying much.
The US is already way way WAY behind in probably the most lucrative markets in the world.
Cell Phone technology and PDA technology.
You will never EVER find anything like Nokia's products or Sharp's Zaurus 760 from a US based company.
ALREADY, as the entire world moves to Linux, we are going to be saddled with a under powered, extremely expensive operating system to run all our "patented" ideas on.
ALREADY these sorts of practices are destroying the Information Technology sector in this country as companies use the law to destroy competition and maintain blatently illegal market share in this country.
We are already way WAY behind in the world in many technology markets, and it will only get worse.
The US is losing all of its tech jobs over seas at a frightening pace because companies can't make products here cheap enough. Half the time because they are funding too many lawyers or the entire board of directors to the CEO are as crooked as a tree branch.
If you live in the US, and you have a large sum of money. Put it in your pillow, a savings account or buy gold or invest outside the country, but for god sake don't put it into anything American.
Just graduated from College with a CompSci degree?
Want a piece of advice?
Move to India. That is were all the programming jobs are and the biggest expansion of salaries will be in the next 10-20 years.
Otherwise go back to school and get your Masters Degree, THEN move to India.
Your chance of getting a programming job is minute to small.
There is no way on earth Windows is going to get anymore secure. In fact, it will only get LESS secure as time goes on.
Why is that you say?
Simple. Windows is a monolithic piece of software.
You say so what and what does that have to do with it?
Plenty.
For starters, anyone who knows anything about Software Engineering, and what came out of the DARPA and TRW research agencies in the 80's/90's, building some of the largest edifaces of software ever concieved is: Less software makes more secure software.
That is, you don't make a software system, more secure, by adding more secure software to it.
Software by its definition is not a discrete mathematical concept, it is quite open ended. Therefore you can never compute, predict, or make software, by definition secure on a Von Neumann computing device.
It is a pipe dream.
(That is why I don't subscribe to the idea that software is patentable, as some say because it is like a physical system, machine. It isn't. Software is a method for producing abstract mathematics. It has no bounds, and all thinking methods in the known universe are employed to make it work. By its definition it is PUBLIC DOMAIN.)
Furthermore, agencies working in secret, building black projects with the tax base of the entire US trillion dollar economic base at thier disposal, have studied these issues with far more resources than any commercial venture can muster, certainly more than Microsoft has.
I find it ironic that even the most declassified basic research that comes out of the published reports on the DARPA software engineering/development organization goes unheeded by so called "Microsoft Press" book publishing "experts".
The only way to make Windows more secure is to start copying the way Unix was concieved of and built, and take a lesson from history.
That is, Linux/BSD/Unix allows you to CUT OUT large sections of software that need not be loaded or used on a machine for a particular purpose by the kernel or OS.
If you could for example, cut out the GUI on Windows, all of the COM/DCOM and.Net services, and leave just a IP Stack and a Pop Mail service on the machine, you go much further in protecting the security of the machine.
But you can't do that. Microsoft is trying to make machines so easy to use that a monkey could operate them.
That is fine, but the world is composed of bigger problems than the monkey and the machine combined.
This is of itself is not bad, software should be easy to use, but software must solve a problem, and I am afraid, business/scientific problems cannot be ALL classified in the same genere as Miss Tilken's and her Mail Merge problem that can be solved with a Dialog Box and OK/CANCEL options in a Wizard.
But this is how Microsoft continues to proceed, building ever more enourmous Operating System Software and applications, taking this philosophy and putting it into thier OS base were it doesn't belong.
The Monkey Philosophy belongs on the application level, along with Miss Tilken's. Not at the OS level.
Effectively, Microsoft is attempting to encode every possible "Enterprise" scenario in its products "Wizards", so that the software makes most of the decisions, and the User just pushes buttons.
Worse, those scenarios not covered in this contrived decision tree, are deemed "enemies of the state". (i.e. products, such as third party tools you load on your XP machine can be viewed by Microsoft as 'something we don't support', call back after you remove the offending software..thank you for your $300 dollars, have a nice day.).
Then many people on slashdot, and Microsoft, I have seen have said "Well, our products save you a ton of time and make things very easy..."
I am sorry, but software and computers, and particularly the problems they attempt to solve...ALL CAN'T BE EASY. Undoubtedly SOME can, but not ALL. For things like Word Processing and SpreadSheet work,
The only people you will here ranting about standards being a annonyance are:
1) People who equate innovative software with proprietary, secret ownership, also know as "Intellectual Property Rights" idiots.
Specifically, propritary software becomes more valuable because you can make it illegal to write software that does the same thing, or even interoperates with it.
2) People who don't like standards LOVE proprietary ones so that they can jack the cost of software that we pay for through the roof.
3) People who don't like standards are usually companies that have a monopoly on the marketplace and find standards a threat because it allows competition to interoperate with thier products and allow consumer choice.
Everyone knows that in the TRUE BUSINESS MODEL, choice is bad for the consumer, right?
4) People who don't like standards:
Bill Gates "Standards Stifle our innovative product EULA's, our software that calls home on your internet connection dime. If only we had a proprietary protocol so you couldn't find out those things. MMmmmm, who was that Attorney General we paid off? Get whoever it is on the phone now! I want to do some innovative goverment bribery!"
Ballmer "That Dancing Monkey Man"
SCO "You made a standard that put us out of business. HOW DARE YOU! Where is our legal products department? We are going to sell you a lawsuit instead! You must buy it I am afraid, its the LAW. Muahahahahahahaha."
5) Finally those who don't like standards are anti business. No standards means tons of proprietary products that don't interoperate together. No interoperating products means no competition in a market sector because it is too expensive to develope for (End legal proceedings, royalties, and of course the kitchen sink which investors find is not a great place to wash thier money down the drain...)
No investors no money to fund great ideas.
Is it any wonder, the major technological innovations occurring today in operating systems development is with Linux.
The only thing Microsoft can do is secretly copy the source code into each XP release, which they do.
You will never know either, because you could go to jail if you found out they were doing it with the screwy IP (Intellectual Property) laws here in the US.
And Finally...
The US economy, specifically, the technology sector is sick.
No I take that back, it isn't sick, it is in its death throws. I am not sure just how many realize that most of the small innovative companies in the US are no longer.
The only thing the US has left is a small bunch of companies, that are too busy "innovating":
Like WorldCom who is putting new innovative techniques into how to change thier name so they can screw more investors over like a cheap whore on the street corner.
Which, by the way, you get more for your money.
Yeah, the tech sector in the US just looks great. I can't wait to see what kind of crappy products are comming out next year from Microsoft, and everyone else in the software industry, currently being run by goons like SCO, Microsoft.
Slashdot Mirror
After developing applications for a wide variety of banking industries it became clear that:
1) The only way to develope software systems, is to proactively secure the systems once they are deployed.
2) To proactively and continuously review and examine such systems, you must have the source code and build tools and access to the hardware engineering requirements of the systems involved.
3) The only known process where this can be achieved is through Open Source.
Closed binary proprietary software is not secure, cannot be MADE secure, is impossible TO secure and with patents and copyrights laws as written it could be quite possible you could be SUED for securing the software yourself.
Security became an extension of the software engineering process for the company I started previously, and it involved reviewing the source code and making changes, performing attacks, etc.
Critical to this process was to have as many eyes and opnions looking at the source code as possible. The more experienced professionals that had a chance to offer advice and opinions on the code, the better and more secure the code became.
An entire portion of the software engineering process cannot even be done with proprietary software, and I personally as a CIO, declared proprietary binary only software sales DOA in this industry 2 years ago.
-Hack
Yes, there are problems.
Primarily, if you are using the "experimental" options in a kernel build.
If you attempt to add experimental options in a kernel that uses a different kernel config file, the results are undefined.
That is, the kernel isn't guaranteed to build correctly with experimental drivers, adding a diffeent config file with experimental options or built in options that need to be configured at compile time, could result in values that are compiled for modules that could cause an oops.
A good example, is the tag settings for a experimental SCSI driver. Usually you can set how many SCSI commands you can have outstanding, for your SCSI bus, before the device demands attention again from the driver.
Say the author put in a value of -1 for the experimental driver in release 2.4.x.
You makde a config for that and put it in a recent release, called 2.4.x+1.
Well, 2.4.x+1 is no longer an experimental release, and -1 when compiled, will be outside the range of the expected values. The Author, expected you to generate a new config, which WOULD have provided the module with the proper info.
So you compile it, and when the final release driver gets a -1, it oops the kernel, trashing your partition.
-Hack
You know, I talk about the state of software with a large number of companies.
...the rest of the presentation, I hope to have up for download soon. I am working out the kinks with the audio make overs for the slide show, but hope to have it all done soon.
All of them clueless, specifically, about what is about to happen to the IT industry/software industry world wide due to the open source movement.
The biggest losers? US and Europe?
Why is that? Continue reading to find out why China is the next 21st century super power, and the US and Europe will be side players for the most part in the later half of this century.
At the moment I give open office presentations, if I can get an audience, to companies and hope to have a multimedia version of it built so you can download it at my website: http://www.aesgi.com
But, here are some of the things I point out:
1) The computing industry, historically required companies and corporations of individuals, so that capital could be organized along with labor to produce software.
This was due to the fact it was incredibly expensive to obtain hardware, and software was a real bitch to build.
You HAD to have a corporate entity, or government backing to build software.
2) As our industry has evolved, hardware has become extremely cheap and powerful...through competition. Software Engineering has allowed tools to progress that enable extremly sophisticated views of the software process to become possible.
This is because hardware has become SO powerful, that and individual now can now rely on whole "environments" to make very smart decisions in constructing software, or at the very least, inquire about contributions to said software from a colleague.
3) The internet has eliminated the need for corporations/institutions to build high quality/very complex software. For very little cost, organizational units that are far superior to what can be bought can be assembled at very little cost using the internet.
Intellectual capital using this method of organization and procurement is superior to what any single corporation can purchase.
4) The corporate mantra/dogma of build software, release software for sale, then add features, then sell, then add features then sell again, is dieing.
Users don't want more features for often than not, when they only use 10% of a program's feature set anyway.
If all you do is write software, and all your revenues come from only sale of that software your business will start to die very soon.
Whole countries are rebelling, due to the fact that software costs are increasing, not because they have to, but because of Western political interests which seek to lock the market up with unfair laws, and to maintain the status quo.
5) Intellectual Property and Copyright law is about to undergo a MAJOR shift.
American companies are enacting Digital Copyright protection mechanisms that are fundamentally unsound to free trade.
Foreign countries outside the western sphere of influence know this, and are drawing up plans to protect their interests.
6) Countries around the world are watching the US and Europe. They are taking steps to prevent western influence of technology in their countries to protect their soverienty.
Specifically, using open source software, so that American political interests, cannot shut down a countries computing/information infrastructure should someone declare a software or patent violation of foreign soil.
7) China and its neighbors, have told the US and its allies that current copyright law and patent law as written is not acceptable to Chinese state interests.
China has taken steps to halt and reverse the progression of American/European Software into their country.
8) The Far East, the worlds most populous nations, are building infrastructure, based on Open Source.
This has a number of very important implications, such as costing American/European companies 1000% more than Asian companies, just to route a packet over a network, or do a simple mail merge with your favorite office suite, to simply storing a file or printing it.
-Hack
Err, No. Don't do this.
Configuration files are NOT compatible accross kernel families.
Do NOT use configuration files for RedHat and then attempt to use them with plain vanilla kernel.org kernels.
you will be VER VERY sorry one of these days if you do...
If you do this, you have a ticking time bomb waiting to go off.
Always create a new configuration file with the vanilla kernel you download and do not mix configuration files across kernel releases.
-Hack
And so it begins...
China has made a few things clear with respect to its rise as a new super power:
1) We will show the world what can be done and take the lead as a nation in engineering, space exploration, and computer science.
They started with the largest engineering project ever concieved..the largest hydro electric power project EVER.
No western nation could duplicate such a project even if we wanted to because of the sheer size, and the use of cheap labor by the Chinese.
China will have clean, cheap energy for 25% of the nations total needs for the next 100 years from this project.
2) It has been rumored, recently in a M5 visit to China, that space exploration plans include a moon base within 15 years of thier first successful manned orbital launch.
It was also made in not so many words, that once they complete this base, we are NOT welcome.
3) China, doesn't want Western IT technology. Especially Windows, or Intel's chip technology in any sort of influence on its internal consumer markets.
It was made clear that Intel's Chip ID technology and the CIA's insistence that back doors be placed into foreign copies of Windows, was not acceptable.
It is forbidden in China to attach any Intel processor based system to the internet with Chip ID technology of anykind.
This new processor was a "call your bluff".
After all, HOW DARE YOU compete with Western technology, you can't possible build anything near as well as we can, so you must accept our processors if you want to do business with us.
It is this direct response to Intel's Digital Rights, chip ID technology as well, that this processor now has been born.
It won't take China long to ramp this processor up to Pentium 4 Xeon quality (3 years at most), using Tainwanese acquired fabrication planets and technologists.
In my view, China can sustain its economic growth internally, due to its population size for the next 50 years, and tell the rest of the world to kiss its ass.
I believe greed, our very own Intellectual Property Rights Laws, and this obsession with Digital Rights Management has locked us out of the only market that will provide long term economic security to the United States.
What is more, I believe China is using these laws against our very own interests in doing business over there.
Coupled with thier own version of RedFlag Linux, once China has its own PC, and own Operating System, THEY WILL DICTATE TERMS 50 years from now to the Western world.
Our business leaders are fools in this country, and while China builds the biggest terrestrial project ever concieved, builds its own Moon Base 20 years from now, the world will do everything it can not to make the Dragon DISPLEASED.
-Hack
Consider what I am doing:
:-)
:0) by:
...etc.
;-)
1) Trying to build a not so average publishing API, with some truly remarkable features in Java, using parallel databases.
http://www.aesgi.com
It works, but it has issues, and what is more, its fast.
2) Talking to companies about Linux, on average I spend about 20 hours a week at corporations discussing Linux...
I have to digest enourmous amounts of data to build these presentations, so the amount of research I do approaches about 20 hours a week.
If your in the Madison area, you ARE on my hit list.
3) I run 6 miles a week, or about 5 hours a week doing something healthy for myself.
4) I sleep 8 hours. Very important.
Plus I eat like a pig, but thats OK because I am blessed with the metablism of a small gerbil or equivalent rodent. 5% body fat! HA HA.
I am the disdain of women everywhere when I go out to eat.
5) I am building broadcom chipset drivers for my 802.11G card for my Linksys access point.
Surprise, you can't get braodcom drivers for Linux.
Probably because Microsft has paid them under penalty of IP to make them specifically for Windows...
ONLY.
But I am reverse engineering them using a windows box.
HA!
Just try and stop my from releasing them soon Billy Boy!!!
6) I have to write proposals for deploying solutions I come up with, thats another 8 hours at least.
7) I am a LLC in the state of WIsconsin so I have to do quarterly tax reports, plus handle 1099's plus, I am considering sponsoring a H1B Visa guy, who is a good friend of mine, so he can become a US citizen. Got to do all that business stuff too...accounting, setting up my own computers, etc.
So, I am CONSTANTLY working, and if I may say so I get it all done on time under budget (mostly
Buy a laptop. I can work anywhere and do.
The ole office/apartment can get opressive, so I go out and work in restraunts, the student UNION, Barns and Nobles or Nicks Bar on state street.
Since I have a sager notebook. I get about 40 minutes of battery life, so I have to pick places I can steal electricty from.
Point is, I work constantly, no rest for the weary...or in my case....the wicked.
-Hack
I am surprised nobody wrote me back on how the SOFTWARE INDUSTRY is thinking the same way as the Biotech industry is in the US.
(i.e. You CAN'T OWN your own software, AND you MUST RENT software from us or it is illegal.)
That way, you never write a piece of software that is of any kind of high quality. This enables you to sell defective software, and continuously charge your customers for defect corrections.
Which, is a relatively new concept, thanks to Microsoft.
-gc
If we didn't have so many lawyers involved in the Biotech community, we would HAVE growable organs NOW, and very much less ethical considerations in who to hack up and why.
:-)
Besides, you may be able to live with a donated organ, but donor patients are soon wishing they were dead with the medical bills which they will have to pay for the rest of thier lives, the enourmous health problems associated with depressed immune systems, and the Frankenstein like chemical cocktails they pump into your body to prevent rejection. Many of these chemicals, are poorly understood how and why they work.
Besides, which brings up another point I frequently make. Why would any private medical company EVER want to cure ANYTHING?
Think about it, it IS NOT profitable to cure people. I mean, if you cold grow someone a liver from their own genes they are cured!
What is the profit in that? I would concentrate on doing better transplants, so I could sell people all kinds of drugs, and tons of extra medical care for a transplanted liver.
Not only that, they have to keep buying the drugs from you for the rest of thier lives!!
If they can't pay they die!
Profit or else, or else if you refuse to pay, you die. What a concept!
I think I will patent the idea.
But since only 1 or 2 companies can work in in this area of growing organs, due to patents, it is going to take a very long time indeed before we see any organs you can grow.
(i.e. Most of the handful working working on this problem are idiots, and pretty slow learners as well, I am afraid.)
Besides, like I said, there is now way a board of directors is going to allow growing organs from ones own tissues....it doesn't make good business sense...for the masses anyway.
Now, if your that very rich person on that board, well, of course you can grow your own organs.
I use to work in Biotech and I can tell you the stuff that is considered in closed rooms, shut from public eyes, would make the slashdot audience roll thier eyes in disbelief.
-Hack
MagiColor 2350 by QMS.
Cost my about $900 at Office Max on sale.
More than you want to spend tis true, but it is a damn good color printer.
-Hack
Those being questionale in my opinion with regards to BK and the Linux Kernel, one based on GPL code and the other not.
:-)
Long term these two cannot exist. Why?
1)
The primary reason why is that companies in what I call the old "established" economies of software manufacturing, insist they have complete control of the software.
2)
The second reason is due to the fact that as we transition, and leave these companies behind, to the new economic reality of "software commoditization", said companies will enact/engage DMCA/Copyright and IP laws to prevent competition or improvements to the process of building the Linux kernel.
1 is obvious. 2, not so obvious. I mean, what would happen if for some reason the Linux kernel couldn't be developed, or not easily developed, on something besides BK?
Furthermore, BK isn't the end all be all of software development methodologies to handle source.
You have to remember, that BK and the company that owns it is basically TELLING YOU this is how we will develop and manage software.
Which brings up my final point:
3)
All software in the proprietary space is being usurped gradually by engineers who are building better software, through a process that continues to evolve.
This process isn't tide to a particular piece of software that cannot easily be changed, to keep innovation of the basic practice of building said software from stopping due to a license restriction.
I think I have to disagree with Linux on my final point. The Linux kernel will soon reach an epiphany, probably around Version 3.0. Which will require some fairly advanced pieces of tools to be developed to keep development, and the community on track.
When we reach that place, we will need tools we can modify to bring the process of building the software, as well as the kernel technology itself, to new levels.
I don't see how that is going to be possible, if we continue to run BK.
If this continues, kernel timelines, and code quality will suffer. Already 2.6 is an immense challenge.
We should be looking at, quite perhaps, something like and Eclipse style plugin, or perhaps a seperate tool for KDevelope, to handle specific kernel versioning, and development issues in a standardized environment.
Right now you have to spend a good 30-40 days of continuous reading, kernel track, source code, and putsing around with building perhaps 1-2 machines to build anything serious for the Linux Kernel.
After 2.6, we will need some sort of IDE to get programmers up to speed faster, so they can learn more quickly, about the kernels innards, etc.
We don't need to rebuild an IDE, specific for kernel programming, we could design Plugins for Eclipse or KDevelope to further that task.
What could be included would be:
1) A current and relevant deprecated code completion aware editor. Perhaps, connected with a version control system that would keep developers aware of changes in the main development trunks.
2) A more complete robotic/remote host plugin for developing/stepping through those nasty sorts of device drivers you pretty much need two machines for: 1) Video Drivers and basically Network drivers, for an adequate building environment.
I mean, given the responses from the BK owners, imagine if Sun told the Eclipse team: Don't you dare build a remote debugger, we will reengineer the protocols the the JDA (Java Debugging Architecture) with ever jdk release so you can't debug code without our software.
You can bet Java code would suck even more than it does so today without an integrated debugging environment.
--
In conclusion, I think these sorts of things will be very difficult to create or build, if the very software that is holding the main development trunks of the kernel, cannot openly participate in the building of a continuously better approach, to creating/debugging/versioning/managing source code for the Linux Kern
and someone has WAY TOO MUCH of it appearently... :-)
But seriously, who cares about Intellivision, I want to run OpenOffice on my 2600 game console.
-Hack
"...suck it you Intel Weenie!"
absolutely classic.
Oh well....I am going to go "suck it" now.
(Working on a device driver for my Radeon 9000 128MB video card....)
-Hack
Absolutely.
.Net platform.
.Net's distributed DCOM/DCOM wrappers (which by the way that is all .net is, is repackaged COM/DCOM via the web) was invented so that you buy MORE Microsoft software.
There is nothing open or standard about Microsoft's
The entire platform wasn't even created to solve a business problem, it was simply created to kill Java.
Well, that and because of
Java was created to solve an industry wide, 30 year problem. How do you turn software from an expense, into an asset?
You do that by making the software independant of hardware, and making it a part of your IP property of your business process.
Which should be perpetual, respond only to a business need, and not a vendors whim.
The reason why Java was invented was to free companies from buying a PeopleSoft implementation for 10 million and then having your investment totally invalidated by someone who buys the company and destroys your investment.
Like Oracle plans to do.
Open Sourcing Java would simply free licensed implementations from being engaged by the certification process.
I am not so sure, that is great for Java. Sun has been extremely magnanimous in managing Java.
Having a bunch of Open Source implementations running around with no means of making sure they all work the same way is not what I would call a positive thing for Java...GPL'ed or otherwise.
Remember, Java write once run everywhere is critical. Perhaps not in the US, but too all my European and Chinese and Japenese customers, who are quickly moving away from anything Microsoft, and everything Linux, it is very very important.
Some sort of release mechnism will have to be developed to insure that Java GPL'ed implementations are not allowed to fork with non standard runtime libs.
-Hack
Well,
:-)
since you are talking about FUTURE, and not now, I would first and foremost use Linux to manage the bandwidth. I assume you haven't made any buying decisions yet?
You can buy propritary heavy lifting equipment such as Cisco gear, and collapse the entire management tier into one switch, but I wouldn't.
It will work fine, BUT you can't reuse Cisco gear as desktop PC's, or even donate them to a school when they get too old.
My way of building things, you can reuse the equipment when its all done.
With that said, I would get a rack of i860 Motherboards, such as SuperMicro's with 6-8 Gigabit Ethernet cards in them and build a BGP based solution.
You get:
1) PC Hardware, commodity items. Easy to replace, easy to repair should something go wrong.
In fact, for some customers I have designed such solutions for, I usually put a spare unit in each rack, on the floor, and just shove it in when I need it.
In any case (4U), each rack would have about 10 machines, with one serving as a switching backbone. I would connect them all with Gigabit Fiber cards into one box. This box forms the QOS SLAVE. We had 6 racks of these, so I had 6 QOS slaves, that then were connected to a Backbone machine that monitored each racks incomming T3 Line, and rerouted the traffic from one rack to another based on how much was comming into each slaves backplane. Each of the 6 slaves was connected to the master, which had Nagios, and the master BGP/ASN lists.
The application was a video feed, from 8 different racks inside the same building, going out AND comming in from various sites. Voice over IP was also used over the same T3 lines.
Insuring enough bandwidth and prioritizing the traffic was key.
Most i860 PCI-Xpress (64Bit PCI) i860 Motherboards, especially the ones from www.supermicro.com can handle 6-8 cards in a chassis. Some manufactures sell external PCI-Xpress chassis and can accomodate up to 16 cards externally.
2) Linux Kernel 2.4.21 coupled with BGP, QOS additions to the kernel and a decent management program like Nagios, will give you all the bells and whistles you get with the big boys.
Except you don't pay millions for such a setup.
3) With such a system, you shold be able to build QOS ratios between BGP ASN's where you can alternatively route to a given pipe through a different rack that has the LEAST amount of activity on it. (i.e. in my case, that was, the least amount of T3 activity, easily monitored of course with Nagios and some simple SNMP scripts which updated the QOS tables, and if required BGP ASN lists on the master machine for the rack.)
You can adaptively route based on QOS signals and change your BGP asn lists to reflect traffic levels in almost realtime.
Been there, done that.
Ironically, my biggest problem with all this, was heat and backup power. 400-500Watt power supplies in a colo, that was originally designed for 4 U chassis with 200-300 watt power supplies, made the place pretty toasty. (78-80 degrees).
It is amazing how much heat the new Xeon processors and motherboards throw off, especially if you MAX OUT all the PCI-Xpress slots in them!
New air conditioning system, and an extra 5K main had to be shunted to the building. Total cost was about 32 Thousand. About 1/4 the cost of the entire project so that was a bummer.
After that the average tempreture dropped to 60 degrees inside!
I loved it, I could watch my pretty Nagios 3D graphs and eat my Diary Queen (one accross the street from the colo) Peanut Buster Parfie while watching billions of bytes go by, at least when I the customer called me to do work at the site.
-Hack
Undiscplined thought is a problem with many people.
If your thoughts tend to race, you need more time to slow down. Do you take the time during the day to focus or practice on making only one thought in your mind?
Like learning new things, your brain can rewire itself to learn how to think in this manner, if it is too "wiley", but it takes time and practice.
Mental concentration, or drugs that alter your neural chemistry, can also affect your intelligence as well.
You may be trading the serene mind, for a duller one, restricted by drugs, with fewer thoughts for one that perhaps with proper practice could be trained to control and utlize all those distractions.
Your body is a machine, and for the most part you have greater control over it than you can possibly imagine.
I would try far eastern techniques for focusing your mind for an hour or two a day for 6 months. If this doesn't help you, then try the drugs.
-Hack
I find incredible, to say the least.
:-)
:-)
:-)
While we are on the subject, I believe Monopolies are good.
No, I didn't lose my mind. I am also not finished yet.
I believe in monopolies, when everyone for example uses the same browser for >90% of the people out there and:
1) Source if freely available under either BSD or GNU.
2) Isn't controlled by any one company or individual.
Then monopolies are great.
Hopefully Linux will become a monopoly one day.
-Hack
I think, several challenges to just materials science has to be overcome.
For example, our current science in engineering, relies on models and previous engineering attempts, to build new structures.
If you want to build a structure, say taller than the sears towers for example, you can do so, by using the Sears Towers as a reference, then building perhaps 10-15% taller.
Historically, we buld a large number of structures, not just buildings, a little bigger at a time. We build planes, a little faster at a time.
That is how our engineering science works. Even when we sent men to the moon, as colossal a task as that was, we took very small steps at a time, and it took decades.
Building something like a Space elevator, in the timeframe (10-20 years) I think is ridiculous given our current engineering science and application of Mathematics/Statics etc.
Just because you have a material than can go hundreds of miles straight up doesn't mean your structure will.
Whole new branches of engineering will have to be invented, as well as new mathematics to make this structure work.
Personally I think the work Stephen Wolfram has done so far in FSM's (Finite State Machines), may offer a clue as to how we can take much bigger steps in the sciences, with much more predictability, in our models, and methods of construction, to make a space Elevator possible.
At the very least his work sheds light on the principles of complexity, and why we take baby steps in everything we do.
Specifically, how can we design systems, when we have no working model, and to build such a model requires an order of magnitude in scale our engineering science, historically, has never had to deal with.
I think, after a century or more of using this material in terrestrial structures, to understand how it works better, we can start thinking about such an elevator system.
But I think it is a safe bet you are not going to live to see one anytime soon, much to the contrary some of these guys at the Space Elevator web site will have you believe.
-Hack
I am not so sure.
:-)
My take on it, and I don't pretend to know both sides:
Some guy wants to be a partner after being at a company, for only 8 months, and only has what 6 customers?
I would say the guy is out of his mind, has no business sense, is unreasonable, or has an ego the size of Bill Gates Estate.
Let me tell you, that is one big ego.
Partner in the classic sense, means you bring something to the table nobody else has contributed to the company, that adds a great deal of economic value to the company.
That normally means something you can't get from hiring an employee.
If he had like 100 customers, or invested a million into the business, perhaps.
Being a partner isn't all what it is cracked up to be either...as the saying goes, becareful what you wish for.
For me, when I was upset with the last place I worked, working 80 hours a week, I left to be my own boss.
Now I get to work 80 hours a week, except I get to yell at myself for stupid things I do.
No wonder half the people I work with think I am nuts all the time, talking to myself.
That boss of mine is not only a whip cracker, he has high standards too!
-Hack
Something I sent to Slashdot Editors a few days ago...
Didn't get any bites though.
I think it would be a good idea, to bring a class action lawsuit against SCO.
I think we should use slashdot as a place to organize such a lawsuit.
For the following reasons:
1) I think technically, this lawsuit given the recent changes in management at SCO, involves fraud. That is, the companies officers know privately they do not have a case against Linux, and are fraudulently misleading thier investors/shareholders too personally enrich themselves with regards to stock price using a lawsuit to falsify product value to said shareholders/stock holders.
This is due to the recent in jump in SCO's stock price. The company simply isn't worth the current stock price, historically and is therefore artificially inflated.
2) Technically, I think, from the perspective of most Linux Kernel developers, including myself and SCO's own development group, that SCO may have abridged GNU code illegally. If this wasn't the case, I don't believe SCO would have continued to sell thier own distro after they committed the lawsuit.
The suit should include full disclosure of all SCO source code. Furthermore, seperate suits should be filed against SCO should GNU software be found in thier kernel.
3) The suit is affecting the industry, consulting firms, companies in real, economically negatively, in a measurable way. Customers are being lost, companies are having to spend money to switch, or consult legal people. This is all because of SCO's suit.
We do not need to wait to the end of this suit, we can file class action suit immediately to get damages/satisfaction.
I also believe that if we ajoin the company officers in #1, we should be able to file a seperate lawsuit against each officer of the company, and not just the company as a whole.
I think, we should use slashdot as a place to:
1) Ask people to generate documentation. Documentation of an official nature, which supports points 1-3. For example if you are a consultant, and you lost a job based on SCO's injunction and public statements, ask the customer to write a letter detailing the loss of business because SCO makes Linux too risky.
2) Internally, if you are working for a company, obtain permission to use corporate Email disclosures for any migration plans away from Linux.
3) Detail any personal damages as a result of not being able to make medical insurance payments, bankruptcies, or personal financial hardship as a result of the loss of business as a direct result of SCO's pending lawsuit and its affect on your salary/business.
I would be happy to help organize my time with regards to this, and would encourage anyone to Email from Slashdot as to how to proceed.
Please post.
Afterward, we can begin the process of selection of legal counsel once we organize.
-Gregory Carter
-CEO
-Applied Engineering Software Group
-gcarter@aesgi.com
And exactly why is this news of worth?
The IP laws fundamentally work against Open Source.
Any company can extract code from an Open Source project, such as Microsoft and then incorporate it into its product.
As such, IP law protects the company from this sort of illegal appropriation because of disclosure rules governing IP law and the DMCA act.
What we need, is something akin to the BSA and SPA. A "tattel-tale" website.
SPA encourages employees to tell on thier companys if they are pirating binaries.
Why don't we have such a website that allows employees to tell on companies that pirate GNU Source Code by incorporating it into thier products, and not contributing the changes back to the community?
After all, do to the enourmous amount of corporate corruption in the US, under the table political manuevers our #1 enemy is doing, there must be a huge number of burned out pissed off Microsoft/ACME employees out there.
We only need one.
Rewards would be part of the legal settlement, should money be awarded.
-Hack
I think the primary problem is that the technology to build and design probes changes too quickly, and affects design.
I always thought that there should be a way, to build a probes navigation and propulsion systems in a standardized whay so that avionics software wouldn't need to change that much.
Sort of a standardized platform if you will for doing solar system exploration.
This platform would consist of a number of parts that would not change, and could be reusable in a number of different configurations for building a probe, depending on what its job was.
Cameras, photometers, spectrometers, and power sources could all be packaged in the same why depending on the probes job.
Every probe that nasa launches is always customized and built around cost and included packages.
I am not so sure that is the best way to go about it as you have to reinvent all the software to manage the probe every time you build one.
Probes should be cheap, produced in high volume, (thousands) and interchangeable.
With a standardized approach, failure rates should come down a bit and costs should be reduced.
-Hack
Here is my opinion. :-)
.Net quickly enough or Java, is through certification programs, I feel sorry for you.
I find a computer science degree exceptionally useful, for the following reasons, IF they meet the criteria:
You attend a University that contains the equipment and staff that will provide you with things you can't get in the private sector, that interest you.
This could be a wide variety of things...for me it was cutting edge research into the areas of vector processing/parallel compiler design.
To that extent I attended UW-Madison, which recieved a grant for 10 Million, in 1991 for a Thinking Machines 128 Processor machine.
I realized that there, I could get access to equipment and more importantly, Doctor James Larus, who had a large amount of work that he published on the topic that I could gain access too. I had a field day for 5 years.
But like you, left to start a company. Now I am back at the UW, 12 years later, finishing what I started.
So, my view on a Computer Science degree, and why you pay for it, is because you gain access to people and pieces of technology that you can't get in the private sector.
That is why I went to UW-Madison, that is why I returned. However, this is not always the case. Most colleges have very little in the areas of cutting edge equipment and "second rate" PhD's, that don't publish much or are not the leaders in thier fields. They are what I call teaching PhD's.
There are TONS of Universities that fit that description. They are still good, but just not AS good. Try and avoid these, but if you can't get into a good program, make the best of it.
Universities, also provide instruction into "foundations in science". What I mean by this is, although the tech you might learn might go out of style, Calculas never will go out of style. You learn calculas, deductive reasoning, and the rigours of the scientific method and research.
In the end, learning how to teach yourself very difficult abstractions and science, analytical thinking, will provide you with the means to avoid what I call "tech scams". The biggest one, is so called certifications in product families companies charge people for being paper experts.
These items Universities teach, don't go "obsolete". Calculas will be taught with the same fundamentals 500 years from now, for example...
Now more on the scams, once you get out of college....
However, I hold just the opposite view on Certifications. (CNE, MCSE, JCP...etc)
Certifications, don't provide any value whatsoever in my opinion. All they teach is extremely short term knowledge, about usually product families with lifetimes, say no more than 6 months or less on the market.
The Cisco, Microsoft, and related Java technology certification programs amount to memorizing product features.
I don't need to pay 10K and sit in a room with an instructor to figure out technology products. I use my own brain, Barnes and Noble's and some independant study after purchasing the product to become knowledgeable.
If they only way you can learn
Much better, in my opinion than some of the instructors who teach the class, I am actually using the product or technology to solve the problem, myself.
I can usually do it FAR FASTER, than purchasing a certification program too. I also feel much more confident. As well, I am not stuck someplace in a motel room for a week away from work either.
So, finally, with my requirements, I would say you are not getting a good deal, minimally, your basically getting a certification level type of knowledge just for a piece of paper.
If you really wanted to though, doesn't sound like your getting much, so I hope your not paying much.
-Hack
What do you mean far behind?
The US is already way way WAY behind in probably the most lucrative markets in the world.
Cell Phone technology and PDA technology.
You will never EVER find anything like Nokia's products or Sharp's Zaurus 760 from a US based company.
ALREADY, as the entire world moves to Linux, we are going to be saddled with a under powered, extremely expensive operating system to run all our "patented" ideas on.
ALREADY these sorts of practices are destroying the Information Technology sector in this country as companies use the law to destroy competition and maintain blatently illegal market share in this country.
We are already way WAY behind in the world in many technology markets, and it will only get worse.
The US is losing all of its tech jobs over seas at a frightening pace because companies can't make products here cheap enough. Half the time because they are funding too many lawyers or the entire board of directors to the CEO are as crooked as a tree branch.
If you live in the US, and you have a large sum of money. Put it in your pillow, a savings account or buy gold or invest outside the country, but for god sake don't put it into anything American.
Just graduated from College with a CompSci degree?
Want a piece of advice?
Move to India. That is were all the programming jobs are and the biggest expansion of salaries will be in the next 10-20 years.
Otherwise go back to school and get your Masters Degree, THEN move to India.
Your chance of getting a programming job is minute to small.
-Hack
There is no way on earth Windows is going to get anymore secure. In fact, it will only get LESS secure as time goes on.
.Net services, and leave just a IP Stack and a Pop Mail service on the machine, you go much further in protecting the security of the machine.
Why is that you say?
Simple. Windows is a monolithic piece of software.
You say so what and what does that have to do with it?
Plenty.
For starters, anyone who knows anything about Software Engineering, and what came out of the DARPA and TRW research agencies in the 80's/90's, building some of the largest edifaces of software ever concieved is: Less software makes more secure software.
That is, you don't make a software system, more secure, by adding more secure software to it.
Software by its definition is not a discrete mathematical concept, it is quite open ended. Therefore you can never compute, predict, or make software, by definition secure on a Von Neumann computing device.
It is a pipe dream.
(That is why I don't subscribe to the idea that software is patentable, as some say because it is like a physical system, machine. It isn't. Software is a method for producing abstract mathematics. It has no bounds, and all thinking methods in the known universe are employed to make it work. By its definition it is PUBLIC DOMAIN.)
Furthermore, agencies working in secret, building black projects with the tax base of the entire US trillion dollar economic base at thier disposal, have studied these issues with far more resources than any commercial venture can muster, certainly more than Microsoft has.
I find it ironic that even the most declassified basic research that comes out of the published reports on the DARPA software engineering/development organization goes unheeded by so called "Microsoft Press" book publishing "experts".
The only way to make Windows more secure is to start copying the way Unix was concieved of and built, and take a lesson from history.
That is, Linux/BSD/Unix allows you to CUT OUT large sections of software that need not be loaded or used on a machine for a particular purpose by the kernel or OS.
If you could for example, cut out the GUI on Windows, all of the COM/DCOM and
But you can't do that. Microsoft is trying to make machines so easy to use that a monkey could operate them.
That is fine, but the world is composed of bigger problems than the monkey and the machine combined.
This is of itself is not bad, software should be easy to use, but software must solve a problem, and I am afraid, business/scientific problems cannot be ALL classified in the same genere as Miss Tilken's and her Mail Merge problem that can be solved with a Dialog Box and OK/CANCEL options in a Wizard.
But this is how Microsoft continues to proceed, building ever more enourmous Operating System Software and applications, taking this philosophy and putting it into thier OS base were it doesn't belong.
The Monkey Philosophy belongs on the application level, along with Miss Tilken's. Not at the OS level.
Effectively, Microsoft is attempting to encode every possible "Enterprise" scenario in its products "Wizards", so that the software makes most of the decisions, and the User just pushes buttons.
Worse, those scenarios not covered in this contrived decision tree, are deemed "enemies of the state". (i.e. products, such as third party tools you load on your XP machine can be viewed by Microsoft as 'something we don't support', call back after you remove the offending software..thank you for your $300 dollars, have a nice day.).
Then many people on slashdot, and Microsoft, I have seen have said "Well, our products save you a ton of time and make things very easy..."
I am sorry, but software and computers, and particularly the problems they attempt to solve...ALL CAN'T BE EASY. Undoubtedly SOME can, but not ALL. For things like Word Processing and SpreadSheet work,
Absolutely Hogwash.
The only people you will here ranting about standards being a annonyance are:
1) People who equate innovative software with proprietary, secret ownership, also know as "Intellectual Property Rights" idiots.
Specifically, propritary software becomes more valuable because you can make it illegal to write software that does the same thing, or even interoperates with it.
2) People who don't like standards LOVE proprietary ones so that they can jack the cost of software that we pay for through the roof.
3) People who don't like standards are usually companies that have a monopoly on the marketplace and find standards a threat because it allows competition to interoperate with thier products and allow consumer choice.
Everyone knows that in the TRUE BUSINESS MODEL, choice is bad for the consumer, right?
4) People who don't like standards:
Bill Gates "Standards Stifle our innovative product EULA's, our software that calls home on your internet connection dime. If only we had a proprietary protocol so you couldn't find out those things. MMmmmm, who was that Attorney General we paid off? Get whoever it is on the phone now! I want to do some innovative goverment bribery!"
Ballmer "That Dancing Monkey Man"
SCO "You made a standard that put us out of business. HOW DARE YOU! Where is our legal products department? We are going to sell you a lawsuit instead! You must buy it I am afraid, its the LAW. Muahahahahahahaha."
5) Finally those who don't like standards are anti business. No standards means tons of proprietary products that don't interoperate together. No interoperating products means no competition in a market sector because it is too expensive to develope for (End legal proceedings, royalties, and of course the kitchen sink which investors find is not a great place to wash thier money down the drain...)
No investors no money to fund great ideas.
Is it any wonder, the major technological innovations occurring today in operating systems development is with Linux.
The only thing Microsoft can do is secretly copy the source code into each XP release, which they do.
You will never know either, because you could go to jail if you found out they were doing it with the screwy IP (Intellectual Property) laws here in the US.
And Finally...
The US economy, specifically, the technology sector is sick.
No I take that back, it isn't sick, it is in its death throws. I am not sure just how many realize that most of the small innovative companies in the US are no longer.
The only thing the US has left is a small bunch of companies, that are too busy "innovating":
Like WorldCom who is putting new innovative techniques into how to change thier name so they can screw more investors over like a cheap whore on the street corner.
Which, by the way, you get more for your money.
Yeah, the tech sector in the US just looks great. I can't wait to see what kind of crappy products are comming out next year from Microsoft, and everyone else in the software industry, currently being run by goons like SCO, Microsoft.
Yeah! American innovation at its finest!
-Hack
2)