Linked by Joshua Boyles on Mon 9th Jun 2003 16:45 UTC
Editorial This entire article is written as a proposal to a coprporation for a new, very unique computing system. Please offer criticism and suggestions to improve the system, and tell me whether you think it could work. What exactly is the "Edge Computing System" And more importantly, why would I want to go to the trouble of developing it? The Edge Computing System is just that, an entire system, not just a new type of computer or new software suite. The Edge is the means by which you can have your personal computer with you at all times.
E-mail Print r 0   · Read More · 60 Comment(s)
Order by: Score:
in windows, there is roaming profile
by Anonymous on Mon 9th Jun 2003 06:21 UTC

in windows, there is roaming profile that serves a similar purpose.

a flash card isn't enough for all type of documents

At work, some guys have two hours' worht of document of
about 4 GB.

The concept is good, however, there needs to be a massive software modification to fit the "standard" and the return on investment is next to zero.

by Greg on Mon 9th Jun 2003 06:23 UTC

You are basically offering Knoppix on miniDVD with some nifty look upgrades and a big flash card (yes, Knoppix already has the ability to save configuration data and the like to flash). The whole hardware thing is entirely unnecessary--nearly every computer has a CD-ROM or DVD-ROM in it these days.

I STRONGLY doubt that the "open source" community will cooperate with you on this. You are basically a company that offers almost no core customizability of the system and piggybacks on the open source development method. Yes you could view the source, but where does it get you? Your DVD is already burned and your "upgrades" will cost money, and you won't be able to control them. Oh, it's easy enough to say "change all your software to look the same, O free software community, and by the way, do a code audit while you're at it so I sell more and don't get in legal trouble." In practice this will never happen. Also, how do you plan to get the security updates on to the disc? Just download them every time? Or will we have to buy a new DVD for each?

Sounds like you need a time machine for the dotcom/venture capital era.

Interesting Idea
by DCMonkey on Mon 9th Jun 2003 06:27 UTC

Interesting Idea but i've had the impression that Compact Flash is a)slow b)not that reliable itself, and c) has limited rewrite capability.

I do worry about HD failure and often wonder how to make it easy for average users (and me) to easily backup and recover. Here is my pie-in-the-sky scheme:

All computers conforming to this standard have one internal drive and two hot swap drive bays with one containing a drive (other drive optional). The internal drive is for the OS and base libraries/files ONLY (as well as temp files and swap). It can always be rewritten via a CD/DVD that comes with the computer and patches/updates to these files are done via the net (or for a small fee you get a new disk every month in the mail).

The first removable drive is your primary data drive and contains everything you hold dear, including programs (I'm not 100% on that part).

The second bay is for either more storage, or your backup drive. Backups are done by copying from one to the other (or with a utility akin to Second Copy 2000 )

If this sounds like the old days of dual floppy drives, data disks, and backup disks, that is kind of the idea. A simple model with physical components users can understand.

Of course the problem is that hard drive can still be physically fragile. Maybe we need to wait for some magical high capacity, high-speed, (practically) infinitely rewritable solid state storage medium.


PS: Of course I can't get this damn firewire HD I bought to format under W2K, So I can't even pretend I have part of this setup today. Grrr.

One more thing.
by DCMonkey on Mon 9th Jun 2003 06:34 UTC

Note that with my Hard Drive scheme, that if all computers of this type are of a standard configuration (probably never happen, Damn you Moore) you could conceivably take you system in for repair and get a loaner while your machine is away, but keep your data with you.

Not original enough
by Zachary on Mon 9th Jun 2003 06:48 UTC

Well, I know this has been said once already and will no doubt get a billion other mentions here, but this varies too little from Knoppix to be revolutionary. Knoppix is on the right track. It is already free and usable and requires no investment from corporations. I can't see this being that much better than Knoppix; especially considering the investment it requires. Regardless, thanks for the interesting article and keep thinking!

Its been said already..
by Zaphrod on Mon 9th Jun 2003 07:04 UTC

The concept is interesting but not that far off what is already available to anyone who wants to use it. The world can't agree on which desktop architecture (Several major and minor types) to use or which OS (umpteen hundred?) but for this to work as you suggest everyone would have to agree on one thing.

keep working...
by Michael on Mon 9th Jun 2003 07:08 UTC

Anything to obviate the need for Micro$oft Windows is a good thing. As others have mentioned, the idea is very similar to what Knoppix already offers.

I believe there are many take-your-computing-brick-with-you ideas in play right now as well. As of yet, many have not made it to market, but will soon.

The portable computing brick has many more interesting upsides than what is essentially portable data.

by wiggly-wiggly on Mon 9th Jun 2003 07:16 UTC

The idea is nice but with some many varities of OS it will be very hard to get the uniformity and consistency required. Even if Linux becomes the standard there are already many incompatibilities between distros (POSIX only goes so far) I very much doubt a the seamless software needed will be possible.

And I can hardly see Apple fan-boys giving up their hardware...

People like their own computers
by Jared White on Mon 9th Jun 2003 07:41 UTC

I'm afraid I don't see much reason to allow people to migrate from computer to computer like this. People like using their own, personal equipment. Buy a notebook and use your computer on the go. Why would I want to use other people's computers when I have my own?


re:The Edge Computing System
by leslie on Mon 9th Jun 2003 07:44 UTC

Sounds a lot like Knoppix, i think some concepts are open for change, how will computing look like in the future, will we still use Hard Disks? Or will we use an intern memory flash disk , that loads the OS in ram, and runs with out any spinning parts? Maybe we need an lighter Operating System too, that can run on less ram then what we are used to now, on low end systems without complains.
I tested knoppix, it works well, but has its hardware issues. And the compressed file system works faster than the uncompressed one, so an uncompressed file system would slow it down.
I sure hope that some changes will take place, good luck with the project.

by Ronald Bos on Mon 9th Jun 2003 07:49 UTC

How are you going to load your or KOffice or whatever-format documents on a Windows or Mac OS machine? Seems the be a pretty big problem, isn't it?

Won't work
by Charlie on Mon 9th Jun 2003 08:08 UTC

Anything that requires the cooperation of so many different entities just won't work.

You're better off looking at something that 'piggybacks' onto other operating systems to give a similar effect.

An excellent example of this is xwt ( that runs on your various operating systems to provide applications that look and feel the same.

Also, I would have though that the way forward would be remote storage of information so your information is available to you from anywhere without having to carry around an annoying disk.

Isn't this What Knoppix is doing already???
by Mystilleef on Mon 9th Jun 2003 08:20 UTC

Geez, I can't foresee anytime in the future where I'll be paying for the Linux kernel, the GNU/Linux OS and free open source softwares. At least not anytime in the near future. Now if all this was for free. It's nice in theory, but lets be pragmatic, secondary storage devices such as hard disks are getting faster, better and bigger. Take Mac's ipod for example. It's a mp3 hard disk that I've slammed on the floor more than once. I even jog around with it. And it is yet to fail me. So the argument about hard disk reliability while noteworthy, is not as terrible as you make it sound.

This idea is very similar to a Linux distro called knoppix. It's basically a cd and a floppy containing your popular Linux desktop environment,user settings and other pertinent Linux files. In fact, I suggest every Linux user has a copy of Knoppix in addition to their primary Linux distro for those times when you want to do things on Linux as opposed to Windows, at work, at relatives place, in a library etc. ;)

by FireBase on Mon 9th Jun 2003 09:01 UTC

1. Yesterday I installed linux on one of my computers to use them like a proxy using squid. Well, I search for a graphical program to configure squid and I didn't find it. Then I opened the configuration file "squid.conf" and I started to understand it. I would like to see an linux distribution with all configuration files in a control panel with each a graphical program with all the same options the .conf file was.
That's the problem of linux. The distributions that exist don't have any interface to that files in a control panel... for now we must edit .conf files and be an expert to undertand it.
When I talk about .conf files I also needed to edit a file in rc.d to start a program when the os boot.

2. One other thing that is important for linux is to be plug and play of devices, and have an option in control panel to install it easily... Yesterday I install the linux operating system without my modem connected. When the linux is installed and working I connected my modem to it, I make a connection in dial up and I choose modem port, don't work! Then I choose the port i don't remember now, but I think ? TTY01 ? and worked, fine:)
I would like to know what's happened if I was a internal modem...

Sun Corp.
by John Blink on Mon 9th Jun 2003 09:03 UTC

This is the strategy that Sun Microsystem is trying to put forward. Well that is what I gathered from a video on CNET.

not likely to happen
by Interfacer on Mon 9th Jun 2003 09:22 UTC

don't get me wrong: this would be a nice thing it it were possible, but there are several reasons that this is a crackpot idea:

-open source enthousiasts want multiple options for everything. while you would want to use a memory stick, i use a flash card, and others use a usb hard drive. so all those interfaces and drivers need to be supported, but as it is open source, people are more interested in hacking and developing then in designing universal interfaces, let alone usability tests.

-people in the states still use gallons, pounds, miles inches, feet and hogsheads, even though congress has long ago decided that SI is the official way to go. could you imagine what has to be done for a universal computer to work? as it is, it is a miracle that you can phone from afghanistan to the usa.

-while you might like windowmanager universal KDE, i think it sucks and want to use universal gnome. unfortunatly my desktop configuration cannot be read on your system because the window manager designers give a rat's ass about interfacing with software not their own. using the clipboard for copy and paste is still very shaky even today.

i bet there are even more reasons, but this will do.


No compelling need for revolution
by jck2000 on Mon 9th Jun 2003 09:38 UTC

I am not sure a compelling enough problem exists for a revolutionary solution.

1. The Internet makes document availability less of an issue. If I am travelling and forget a document, I can have it e-mailed to me. If I am thinking ahead but do not wish to carry physical media, I can put it up on a ftp or web server. A variety of remote access solutions exist.

2. Web access is getting easier and easier (for instance, WiFi).

3. MS's monopoly status, near-universal HTML and PDF readers and evolving MS-compatability and open document format efforts (the last of which I think is the most important on a long-term basis) make viewer/editor availability less of an issue. Java-on-the-client can also be used to great advantage here.

4. Notebooks are getting lighter, cheaper and more powerful everyday. (In terms of size, they have already reached the physical limits imposed by human hands and human eyes.

5. For use within closed environments (company, school), a variety of thin client solutions are available. I have got to believe that thin client solutions for closed environments I believe are going to be a big growth area -- and might be an area where small Linux-based companies can compete with Sun and the like.

6. With 24/7 network availability, thin clients need only passwords rather than physical dongles or media.

You go too far; try for open formats instead
by Vanders on Mon 9th Jun 2003 09:49 UTC

Hmm, so you've basically described a thin client machine with a CF slot. I think we've tried this before.

While your idea has some merit I think you've gone a little too far. You cannot control the hardware, so just forger about that. While you may not need anything more than your 2.4Gb of storage, or a GeForce 2 card in your machine, someone else may need half a Tb and want a Radeon 9700 in theirs. Not all computer users are equal and they will resist any attempts to make them so.

The same applies to software! You will never get people to standardise on Linux. Even Linux cannot standardise on one distribution. Your examples give choices at every turn; if you're standardising, why not just choose one desktop enviroment, one email client, one web browser, one office suite etc? I suspect you already know the answer; people like choice!

I think what you should really concentrate on is open data formats. Many of the problems you are trying to address would not be a problem if we had a well supported standard office document format, or a standardised way to store basic user preferences. Not only would this make cross platform interoperability a lot easier, you wouldn't have to worry about wether the computer in question was running KOffice or OpenOffice, or Linux or Syllable or Windows. The user could plug their CF/USB Memory Stick/Whatever into the system and it would just work. This is far more likely to be succesful in the long term, and you don't have to restrict anybodies choice to do it.

Internet storage
by Tyr on Mon 9th Jun 2003 10:00 UTC

I can see situations where this kind of system might be handy. If a system was included that allowed easy transparent access to files over the internet (internet storage mounted as a local directory so people don't notice the difference) you wouldn't have the limited storage problem and it would make the system more attractive.

Already done by IBM
by Simon Porter on Mon 9th Jun 2003 10:26 UTC

Didn't IBM already make a prototype of this? It was a tiny computer that you could carry with you and you could plug it into a base station and carry around all your documents etc with you.

Seems pointless
by nonamenobody on Mon 9th Jun 2003 10:26 UTC

I don't really see the point to this when you can already get:
* Memory cards which can store upto a GB (maybe more)
* USB/Firewire harddisks
* Laptops
* and portable computers with a 2.5 inch HDD, but no monitor, keyboard mouse etc. (they plugin just like a regular tower would), which will fit in you pocket (they are little bigger that a 3.5 inch HDD).

by uid_zer0 on Mon 9th Jun 2003 11:18 UTC

Which email client do you want to use (if you're not sure, you can choose more than one, then choose one later)?

* Kmail
* Evolution
* Fetchmail

* Mozilla Mail
* I don't need this capability

Hey where's my favourite email client... mplayer?!

Buy a Palm insted
by J. Dahlberg on Mon 9th Jun 2003 11:27 UTC

Both Palm and PocketPC has alot of similar futures that this concept struggle after. Now when Palm have Bluetooth projectors can be used directly trough PalmOS. This future is very good for presenations.

HEY - DCMonkey!
by Havoc on Mon 9th Jun 2003 11:44 UTC

DCMonkey - You were having problems formatting a drive under W2K? I just went through that with an 80gb external USB drive.

The secret is that under w2k you cannot have a FAT32 partition more than 32gb. NTFS does not have this limitation.

Also, be aware that FAT32 has a 2gb file size limit.

Excuse for my english, I'm italian.
Well, why not use two 1Gb CompactFlash card? One instead of the mini-DVD and another for documents and configuration?
I think your Idea will work with a VIA EPIA-M10000 a C3 1Ghz Processor, 2 514mb DDR modules, 2 1Gb CompactFlash card on parallel ATA (why not 3 or 4? You have another IDE channel...), hard disk on firewire and a Source-Based distribution (Gentoo or Sourcemage...)
A Revolution? I think not, an experiment much expensive (compactflash are too expensive!).
Gianluigi Ravviso

RE: Already done by IBM
by Tingo on Mon 9th Jun 2003 12:11 UTC

Yes, IBM did make a prototype. Now">Antelope sells it as the Mobile Computer Core, or <a href="">MCC for short.

Network Computer (dumb terminal), without the network
by chrisb on Mon 9th Jun 2003 12:11 UTC

Sounds like switching from the von Neumann architecture to the Harvard Architechure. The big problem being what if you want to use some kind of new application that isn't included in the standard Edge Computing disc, suddenly you have to switch right back to the von Neumann architecture for the extra flexability it provides. Your dumping alot of flexability for a minor improvement that doesn't matter most of the time. How often do you actually move between machines, really? At home never. At work very very rarely. Even the sales people going presentations to clients spend more time in the office, on one machine, than in other peoples offices pitching.

Nor can you rely on the OSS movement as your edge computing workstation will be acting much more like a Network Computer style dump terminal, the only real difference is the OS will be booting from a DVD (worse than a dumb terminal as it'll be harder to upgrade and maintain), and your data will be carried with you on the Flash card. Not a hackers (in the classical sence of hacker) machine, so they simply won't use it and you waon't e able to build a big enough comunity to make it take off.

What about simply using VNC?
by Yannick Koehler on Mon 9th Jun 2003 12:55 UTC

Seems to me that a minimal disk or boot device with a VPN connection and a VPN Client would provide so much more for so much less and work today.
The problem with having my personal data/info on a flash card and putting that flash card on somebody else reader
which I have no trust in is a big one.

Would you trust that computer at the hotel to not make
backup of your email or source code you are programming?

I believe that unless the data I input is directly written
to my memory storage without the hardware I don't trust having seen it, that the system is doomed.

Even VNC doesn't offer this kind of protection, so if you're not going
to resolve this, then use VNC has it does that today.

Good Job!
by Jay on Mon 9th Jun 2003 13:19 UTC

Joshua, I commend you for thinking and, in fact, for thinking something all the way through. I can see this working under certain circumstances. Let's scale it down and say a corporation uses the Edge sysem. It has only Edge hardware and the software you mentioned. This could work very well I think. LOL, the only part that made me wonder was the huge amount of software. That would be the opposite of the simplified hardware. A corporation would probably want to use one email client, one office suite, etc. I mean, it doesn't have to be that way, but there could be some simplification in areas like that which are universal and everyone in the corporation uses.

As someone else said, keep thinking and thanks for the great article!

the Edge model
by moataz on Mon 9th Jun 2003 13:22 UTC

Your idea isn't bad, and even though Knoppix is already doing something quite similar, there is always room for competition because Knoppix doesn't really have much yet.

The only problem I see with your idea is that of pricing. You will probably not make the kind of money needed by selling the Edge System for such a low price. I'm not saying that you should increase the price of the OS to above $100, because that may mean losing market share. What you should do however is to think of a strategy that will make you money, whether from Tech Support, or through the sale of proprietory software made specifically for the OS. Also, you should charge companies like HP and Dell on a per-system basis instead of a "purchase once use forever" model. This could increase revenues for you by hundreds of millions.
Another idea is to get into the hardware business of developing the miniDVD system yourself and then sell the entire Edge System as a complete package under you own brand. This approach will obviously be more costly and time consuming, but the rewards will be much greater.

Sun Smartcard
by Shawn D on Mon 9th Jun 2003 13:43 UTC

Idea sounds good but there is already a good alternative, Sun Smartcard System. Now instead of going and creating a whole new system why dont you just extend the currently localized sun user database, and replicate across multiple hosts in different locations, so as to person A in california, can come to NYC and be able to log in. And sun hardware is very reliable...speaking from experience.

Prior Art
by James on Mon 9th Jun 2003 13:49 UTC

Sun Microsystems has had this model for years. They use it internally whereby an employee does not have a perm office of PC. They get to work & insert the smart card which authenicates them & reloads their last saved desktop state.
See for more info.

Roaming Desktop
by Mike Hearn on Mon 9th Jun 2003 14:14 UTC

Me and a few other people discussed the possibility of adding roaming desktops over the 'net to Linux. It's not hard, mostly relies upon efficient caching remote filing systems, along with a decent mirror network and good security.

Then you just patch GDM (and KDM), add some code to PAM and now you can enter as your login address and get your desktop pulled up wherever you go.

A bit more achievable methinks ;)

This is OS News?
by Mr. Cancelled on Mon 9th Jun 2003 14:39 UTC

It sounds more like someone trying to get others to build his dream system for him based on a 4 page essay.

Why others would suddenly stop development on an existing OS that has potential to build this "dream OS" is beyond me.

Perhaps you should begin working on it yourself before wasting peoples time with such proposals, then others might be able to better see the benefit of such an endeavor.

English proper not
by Joe on Mon 9th Jun 2003 15:51 UTC

"This would make the computer extremely fast, and if you don't have that huge of an amount of ram, there"

that huge of an amount....Wow, this doesn't scream "stream of consciousness". Must learn to edit...edit, ma boy.

Lots of effort, little payoff.
by Bored Troll on Mon 9th Jun 2003 15:55 UTC

Let's start from the beginning. Your "data is the computer" model is interesting, but how does it vary from the portable hard drives available today? For one, the entire OS is on the portable for consistency's sake. But remember, the station should have local driver space, so that hardware expansion doesn't require a full OS upgrade for every OS out there. So now you are mingling userOS and stationOS, an interesting idea.

User data is stored on the hardware, and the hardware is physically carried with the user. As it sounds, you are describing a PDA, HDD MP3 player, flash drive, or any of a number of types of devices. Furthermore, you are describing a piece of hardware that can store less than 5 GB of data. I have more personal photographs than that, let alone music tracks, video files, old papers, scans of magazines, etc. This is not an acceptable amount: put it on either a portable or a microdrive, or nobody will want it. A backup station should be cheap enough to implement, and necessary for any device that is carried in the user's pocket at all times. Do you think nobody loses their cellphone?

You're making an inherintly single-user, single station OS from the ground up and you're starting with Linux / X? If you want it to be easy to use with clean programs you need to dump Linux. It's a great server OS and a decent multiuser desktop OS, but as far too overengineered to be a top-notch singleuser OS. Align with a Be group, the Amiga group, or design your own OS. But no more jimmying Linux into every corner imaginable just because it can be done.

To Summarize

What your describing is a portable storage system, that contains Data, Programs, and OS. Many current programs can already be installed on removable HDD's (and many more could be modified to, if the Programmers believed in that kind of thing). So all you are adding is a user-controlled OS. Somehow, that's just not compelling.

The problem in slinging data around OS's is compatibility. The problem with creating a radical new hardware standard is... compability. Good luck Sissyphus.

I see two better routes. Route 1: push Motherboard makers to allow booting from the USB port. Exactly what you want, achieved instantly with Knoppix and a cheap Matrox drive. Route 2: Bios and TCP/IP based remote HDD mounting, with local RAM caching. Universally accessable, and will be done about the same time the Telcos finish laying out last-mile fiber.

Your route: Lots of effort, little payoff.

Security and falling prices
by Russell Coker on Mon 9th Jun 2003 16:29 UTC

The first problem I see with this is the issue of security. I am not going to stick a storage device containing my personal data in someone else's computer, there's too much risk of trojans. Technologies such as Palladium/TCPA have the potential to offer some solutions to this problem, if you trust them (and I don't).

Computer prices are constantly falling. Even by Australian standards hand-held computers such as the iPaQ are becoming quite affordable, and are more usable for editing files on the go (I've written a magazine article on an iPaQ). Laptops are constantly getting cheaper, lighter and more powerful. 10 years ago it was possible to purchase a new laptop and a new car and have the car be cheaper! Now there are schemes in place to offer cheap laptops in countries such as Thailand which will have significant flow-on affects to first-world countries. I anticipate that in the near future a cheap laptop will cost the same as a good bottle of Whiskey or Champagne!

I expect laptop and hand-held computers to become so cheap that big corporations will regard them as disposable items rather than as assets. When they are 12 months old they will be disposed of if they have anything slightly wrong with them or if there is a better model. If the issues of doing a quick install on a laptop with full hardware support are adequately solved (NB there's some money to be made in this area) then corporations may make it a policy to just give all employees a new laptop every year and let them keep it at the end of the year. This will flood the market with hardly-used second-hand laptops which will make prices really nice for those of us who don't need the latest laptop to run Linux.

I expect falling prices of laptops and hand-held computers to kill the market for the current Internet cafe business model. I expect that future Internet cafe's will have a wireless hub and a large number of power sockets, and no other high-tech facilities. Incidentally I used to part-own and run Australia's longest-running (at the time) Internet cafe, so I have some confidence in my ability to predict the Internet cafe market.

Intel Personal Server
by Brian Luft on Mon 9th Jun 2003 16:39 UTC

Intel has been developing a similar idea.

Re: Linux and Control Panel...
by technodev on Mon 9th Jun 2003 16:47 UTC

This is not a very feasable solution for feature rich programs. You're talking about a graphical representation of a config file which quite possibly has thousands of options, and in terms of applications like Apache, has a modular design to add even MORE options.

The reason I don't feel comfortable within Windows in the first place has to do with not having enough power over it's configuration. Sure there's the registry, but that thing is less helpful in what does what than Linux config files, and finding a complete documentation of it online does not appear to be possible.

Limited functionality or extendability of programs simply because there "should be" a graphical representation of a config file seems bad. And what's the difference from changing values in a config file and in a graphical represenation of it in the end... you still need to know what the values do. Instead of hitting save at the end you hit ok.

I'm sorry but this idea just seems bad. The downsides outweigh the good sides, and it really offers no significant benefits.

Closed system.
by Will on Mon 9th Jun 2003 16:51 UTC

Fine idea.

It won't work.

Here's the problem.

The problem is simply that it's a closed system. It has to be, otherwise it doesn't work.

By closed system I mean simply that there is a one entity at the top of the totem pole that controls the entire thing.

In order for the utopia of total portablility to be fruitful, you have to control any aspect that affects that portability. One significant aspect that affects that is simply features in the software. If I need to update that presentation at the hotel, the hotels software better be completely compatable with my presentation data.

Could it be? Sure, but recall that the primary way that developers distinguish their software is through features, program flow and integration. If everything is based on a Lowest Common Denominator feature set, then there is little motivation to improve the toolset? And who decides when an feature can be added in later?

Simple example with the presentation software. Consider that the software that developed the presentation may have had some snappy transistion that the hotels software didn't support, or even the final desitinations software.

Today, what you want pretty much exists. With something like the Java Citrix client running in a web browser, you can practically access a rich centralized computing experience from just about anywhere. It wouldn't surprise me if you could do it from those airport internet terminals.

As long as you have an open system, you're going to have incompatabilities. And you must have an open system to attract developers and users.

Microsoft is in the perfect position to really break through on something like this. Selling an annual subscription to office applications accessible through terminal services, and having a fast, downloadable, lightweight client so when someone shows up at a Internet Cafe, a friends office, a Kinkos, or whatever, they can simply install the client and connect to a MS server that had all of their data and applications, from anywhere on the planet.

Most people don't have this need though. They're content managing the entire process themselves.

Re: This is OS News?
by Jay on Mon 9th Jun 2003 17:00 UTC

Mr. Canclled, you are way out of line. An 18 year old young person has an idea and your remark is all you can come up with? I'm afraid your remark is a reflection on you and not on Joshua.

FireBase: Webmin
by Joe on Mon 9th Jun 2003 18:16 UTC has what you need.

by Pieter on Mon 9th Jun 2003 18:23 UTC

Wow ! I just got a big deja-vu reading this article!

hard drive failure?!
by mini-me on Mon 9th Jun 2003 18:23 UTC

What world do you come from? ;)
I have had 2 macs in the last 10 years at home, never had a hard drive failure!

I have been working as a computer person for the last 5 years in a university and I have probably seen a hard drive fail only once, and this is in a duration of 5 years, and over 1000 computers ;) ---

by dwilson on Mon 9th Jun 2003 18:35 UTC

The cost of buying enough flash memory to store my documents would be tremendous. When you consider television shows recorded with my WinTV card, music rips, dvd-rips, songs I've recorded (all in wav format), you are looking at a lot of data. People need to learn to backup irreplacable data and settle for the fact that if they don't back it up, there is a good chance they will use it.

Until we get past hard drives with moving parts they will always have an obscenely high failure rate (compared to other computer components).

Hey - Havoc
by DCMonkey on Mon 9th Jun 2003 18:42 UTC

Thanks for the tip but I was already trying to format NTFS. Ther is a known bug that sounds like the problem I had (something keeps a handle open on the new volume, thus disallowing the format), but the stated workaround didn't work and the hotfix was supposedly included in W2K SP3, which I already had installed. I ended up plugging it into my XP Pro machine and formatting it there, but now have some other problem where the format didn't complete at 99%. I've got a few more days before the return period runs out to tinker with it.

As for my dual removable HD idea, it occurs to me (duh!) that a consumer friendly RAID mirroring setup with removable drives would likely be adequate for what I had in mind.

Lots of good ideas here all around.

Re: Tehnodev, Joe
by FireBase on Mon 9th Jun 2003 18:51 UTC

Tehnodev, I read the file of squid.conf when I start to configure proxy, and I understand it, but in graphical mode is a fast way to do the same thing... I'm a programmer and I know that an application that have (edits, checkbox's, tabs, etc...) is more easy to use... you don't need to be an expert. And you can have help files in it (one suggestion is to put the comments of that file .conf) in the graphical application.

I think that if one of the distributions make graphical interfaces to linux .conf files in a control panel, they will be sucessfully... Beginners users want graphical interfaces, don't want text modes to configure things, that's the past.

You will agree with me, that if you have to choose in a text mode and a graphical mode, you will choose graphical mode to do the same thing...

Thanks joe:) webmin is a good solution to interface config files.

Great way to erase a person's memory
by Berend de Boer on Mon 9th Jun 2003 20:16 UTC

Just put his stick into any computer and wait 30s. Voila. Life gone.

This just one of the uses for NomadBIOS
by Jacob on Mon 9th Jun 2003 22:00 UTC

nomadbios allows you to hot-migrate your OS with all running applications. Write an USB-driver for L4 and nomadbios, and be done with it :-)

re: everybody
by Josh on Tue 10th Jun 2003 04:04 UTC

Hi, this is the author of the article, and before I get started answering all the questions, I have a couple of things I'd like to say.

First: I did not post this article with the hope that people would send me money to develop it (I couldn't even if they sent me money). I posted it because is was a failed project. If it really had promise I would be off making millions of dollars. I wrote this a few months ago when I had some time on my hands for fun, but decided to post it on OSNews just to see peoples responses.

Second: Although Knoppix was one of my main influences (the whole bootable CD thing is entirely from there) the "persistant home directory" had not yet been implemented, and wasn't until about a month after I finished with this project. I came up with that idea on my own, although I'm by no means claiming I invented it, because looking back at knoppix from the past it seems that many people did this. I was very excited when the knoppix home directory came out, because I used it just like my idea, having a knoppix disc at home and work and shuttling my stuff back and forth. It was great.

Anyway, on to the critisism. First, I wonder if anonymous would post information on the roaming profile, because I've never heard it.

Greg: the source would be viewable, and downloadable, for whatever changes you would like to make. The changing of software to follow a common GUI and the code audit would be done by the company producing the discs, not the open source community (although, as I said, the source could be incoporpated into the common code by downloading it if it was an improvement. Also as I stated in the article, security updates would be downloaded at boot time, and incorporated into the running code (kind of like in Cisco's router OS, how you have the running config and saved config). I also view it more as symbiosis, not "piggy backing."

DCMonkey: You should get a RAID controller and mirror your main drive, or maybe make two mirrors just to be sure.

Zachary, Micheal, Leslie, Mystilleef: It was original at one point though . . . (sigh), and Leslie is right about trimming down current OSs.

wiggly-wiggly: what kind of name is that? Oh yeah, and you're right about the apples.

Jared: I think this idea would be excellent for students though, who have library computers, home computers, classroom computers, etc.

Ronald: Downloadable would be a program for windows or Mac that can read the filesystem and convert office types.

Charlie: That's true about so many people cooperating, but I tried to avoid depending on the internet for transfering of important and personal files because to me there is too much risk involved.

John: Could you link to the article about Sun?

Interfacer: You can get a six in one card reader that fits in a floppy bay and reads pretty much everything, I agree with the hosheads point, and knoppix uses several windowmanagers and they run fine (well, maybe fine it too strong, tolerable) together.

jck2000: I don't really trust the internet for private/sensitive stuff that well, although it's okay for some things. Your other points are very valid.

Vanderas: Although I think open formats would be better than this probably, I doubt Microsoft really wants to cooperate with Linux. Hardware it autodetected like in Knoppix, so you don't need the same hardware, whatever hardware you have will work.

uid_zer0: Sorry about that, apparently I came to the mistaken conclusion that fetchmail had to do with e-mail. What is it's purpose? I just threw it on there because it had mail in the title to give me more options.

Gianluigi: Yeah, compact flash is way to expensive for that.

chrisb: Wow, I have no idea what you just said. von Neumann? I fell dumb.

Jay: Thanks a lot.

Shawn, James: Thanks for the link James, that's actually really cool sounding.

Mike Hearn: Probably more acheivable, but less secure, I think. It's a trade off you'd have to choose for yourself.

Mr. Cancelled: I didn't say anything about stopping any OS. In fact, I don't even want people to build my OS because I see the obvious flaws in it. If, however, someone can take one idea from mine and turn it into something better, then it's worth posting it.

Joe: I was drunk.
No, just kidding, I didn't do a huge of an amount of proof reading.

mini-me: macs are supposed to be reliable. Last month at the (windows) pc shop I work at we had at least ten failures of hard drives (and most people didn't back up at all).

Berend: That's why you backup.

Anyway, as I said, it was just an idea, and one I already thought was failed. I do think burning knoppix on a mini dvd would be cool (I've got a friend with a DVD-Burner who I'm going to get to do this to try it out), and I think in installation of some of the OSs with many programs a choosing default type thing would be cool, with detailed descriptions of programs, not: a mail client. Thanks for the responses everyone, even the guys who thought I was a retard.

Re: Design Ideas
by Ben Atkin on Tue 10th Jun 2003 05:58 UTC

This is another example of a great idea that is hard to implement even though all of the code is available and its core concepts are quite simple. This goes to show that there is a long way to go in managing programs and data and in designing programs so parts can be changed more easily.

The reason I got into open source in the first place was because all of the components were there, but integration is still really tough, even with access to the source code. I would like to see more projects done in languages where programmers will more artfully design the programs (SCHEME!) rather than always use an "Object-Oriented Approach", which in my opinion, is great for some designs but bad for others. I think that the statement "The goal of this fuction is to transform a certain type input into a certain desired output" is just as valid, if not more, in computer science, as the statement "An object is an entity with responsibilities", though lately the focus has been on the latter.

Just my 2 cents.

where in oregon?
by Elijah Buck on Tue 10th Jun 2003 06:25 UTC

Where are you in Oregon? 18, huh. So, you just graduated? two more days for me (well, till summer. I'm a junior)

Nice to see ...
by clasqm on Tue 10th Jun 2003 13:00 UTC

... That 18-year olds are still figuring out how to save the world! It may not exactly be world peace, but if you don't think big when you're young, when are you going to do it?

Don't let these old hacks grind you down, kid. You'll have plenty of time to be just as negative and jaded as they are.

edge computing issues
by anonymous on Tue 10th Jun 2003 14:21 UTC

1. Flash is too expensive and too small.
2. Flash does have moving parts - the connector on your edge computer, and the pins on the flash card. These WILL wear out.
3. Flash can and does wear out as it has a limited number of writes. You can burn out a flash card by repeatedly overwriting it. A good example of this is taking a Linear Flash PCMCIA card used heavily in a Cisco router/switch and attempting to use it in a Newton. You'll find your data vanishing.
4. KDE and GNOME are desktop environments, not window managers. They in turn call Window Managers - if I recall correctly (and I might not), at least GNOME 1.x calls enlightenment or sawfish, etc. You can use other "lighter" window managers.

Please do some basic reasearch before posting bull droppings. I'll not add other comments about how you won't be able to get all the manufacturers to work together and agree on a standard. It's already been said.

i think its a good idea
by james on Wed 11th Jun 2003 03:40 UTC

i am probably the least "tech savy" person to post about this but i think that helps me here.

this is a great idea taken too far. obviously the current HD model is limited and flawed, lotta parts moving fast, etc. what is the fastest type of memory we have? system ram can transfer gigabytes per second if im not mistaken.

the problem is that memory deletes itself, but i dont see why this cant be a whole mini-mb with a "bios" where u put any os you want in rom, run it on any computer that is compatible with all the card and the software you put on it, and do the obvious thing that Be did years ago:
just put the drivers "in the box" and if it works it works if it dont it dont.
obviously you could should be able to "flash" something and update it as needed, but we can do this already today.

the only thing that needs to be invented is a more efficient way of storing "optical redundant ram data."

Nice to see too...
by bbrv on Thu 12th Jun 2003 04:25 UTC

Have to agree with clasqm...Keep Thinking Joshua! There is NO failure in this. Here is our two cents:

Evolution occurs in three ways:

1. The strong get stronger
2. The weak die
3. Mutation occurs

The act or process of being altered or changed in genetics involves the change of the DNA sequence within a gene or chromosome in an organism that results in the creation of a new character or trait not found in the parent. Lets look to the computing world you described as a strand of DNA in the day-to-day *life* of people -- we just need to *arrange* things...;-)

The things people do will not change in the normal course of events, but how they are done will if the *new* way is arranged better, more conveniently, cheaper, easier, anywhere, anytime, etc. The key: do the same things better! Just ask your parents about telephones that actually *dialed* and had these funny swiggly cords between the handset and the telephone -- image uncordless and unhandsfree!

Wow, now that's a stretch! ;-)

Raquel and Bill

A Better Solution
by Neil Hegarty on Thu 12th Jun 2003 11:28 UTC

On my home machine I have around 90GB of data and less than 20GB of software! My data requirements are getting larger. A better solution would be to run an application that would do the following:-

In the background,
Mirror my entire system to several other locations on the internet.
Mirror other users' systems on my machine.
Supply a key/card to each registered user.

This should be done without any identifiable information being disclosed to any party. If your system breaks down, one could enter your key and your system could be rebuilt.

With broadband access (albeit an over hyped service) a system could be rebuilt in less than 24 hours.

There are issues, few if any are insurmountable. The biggest problem may be the reluctance of ISP's to allow such a large transfer within 24 hours.

Interesting but.
by Philip Streck on Thu 12th Jun 2003 18:32 UTC

I think one of the biggest problem here is using flash memory. Personally my home directory has 25GB of data in it and its just going to keep growing. Most people are going to have multiple gigs of data and flash memory just wont cut it because as the flash cards get bigger so will people's home directorys. Multimedia files are getting to be higher quality everyday and the sizes of the files grow with them.

IMHO it would be better to store the home directories on the internet with an automated check in/out procedure. Of course there are security issues to be looked at as well massive bandwidth and server potentional. At least this would take the responsibility of keeping a backup to a data center and not a user. Plus what about when a user loses his flash card, or has it in his pocket and it rains..

Thats just my 2 cents tho.

On being a sadder but wiser wet blanket
by Joe Nelson on Thu 12th Jun 2003 19:14 UTC

While I am happy to see such enthusiasm on your part, Josh, I think it's misplaced. Most of what I'm about to say is hypocritical because when I was 18 I channeled my energy in just the same type of direction; but hypocrits' advice can still be correct.

I'd suggest focusing all that energy and creativity to learning. Of course, embarking on a project such as the Edge Computing System will certainly teach you things, but the lesson would probably be something like, "Wow, when I'm trying to create a system that does so many good things for so many people in so many's hard and full of accidental complexity!"

I, too, wanted to make a modular swiss army knife that would be useful to everyone. I called mine Negatron. Maybe these grandiose projects are a fact of life for us young people who really love the power of programming. All I know is that, as months passed, I became more and more aware of my ignorance. I began searching to see what others have thought and written about computer science.

My Dijkstra-like conclusion, so far, is that strict mathematical thinking needs to be applied to this novelty of computing. So I've been focusing my energy on learning Predicate Calculus and Lambda Calculus. In the process, I'm discovering just how smart other people are! Initially, the enormity of what has already been created depressed me. Now it just makes me glad! I don't have to work to rediscover what giants like Church, Turing, and Goedel have already pioneered. After learning this stuff I'll be free to more powerfully accomplish my own dreams.

Not to say that my particular area of interest -- logic and proof theory -- will be what you pick, but just pick something! Become educated, become strong.

by AdamW on Thu 12th Jun 2003 21:02 UTC

it's a thin client, kthxbye.