I'm going to take the Stallman position here. People SHOULD build, perfect and extend these things because - eventually - they will contribute to keeping information free.
I agree with you about consumer viability for these PVR applications. But your position can be used to dissuade home-built versions of a lot of things. Getting them off the ground is part of the fun, insanity and adventure of geekdom. There ARE success stories in that realm.
Now, I don't own a TV, so I come already doubting the value of such information, but that's another story. After all your "saving of money" you're going to watch what drivel?;)
Lower staffing expenses Counterpoint: The methods for administering Linux boxen must be measured in several ways:
Cost to install and initially setup OS for intergration into an environment, including security
Cost build and configure a default user account
Cost to write/configure health monitoring scripts for indication of adverse events
Backup/Restore capabilities and costs
This is only to scratch the surface of admin. In detail, a Linux admin has many capabilities transferrable from older Unix-style OS's. Windows OS's, OTOH, have routinely changed their server administration methods, both through updates and through versions. NT/2000/XP each have delicate nuiances that must be published. This doesn't mention services like ActiveDirectory etc that can radically change a administration model.
More efficiency per dollar Counterpoint: Not counting continuous uptime, Microsoft and Linux can each be tuned to comparable speeds for dedicated servers (mail, file, print, web). This immediately clouds when comparing details since scripting hosts, programming style, cache arrangements and developer expertise may be vastly different. However, given that most tools for Linux are also free, Microsoft must admit the costs for not only the OS, but all their server offerings..NET development, SQLServer, IIS, Exchange servers, etc all come with an ongoing license fee (License 6 model). Also comparing ramp-up time for each of these tools may involve training materials which MS does not give away, versus Linux information flow online and in published info/man pages the costs are stacked highly against Microsoft.
NOTE: Counting continuous uptime, MS loses by such a overwhelming amount that it makes the above counterpoint moot. Cycling a Microsoft box because of a update (forced or not) - across all installed packages - is not uncommon. Keeping the Microsoft tool compatability across these installations is also a hit or miss operation with a poor historical record.
Reduced development costs Counterpoint: Installing all tools from a end-user Linux distro is under the "free" cost of the OS itself. The tools from Microsoft, best bought in a corporate level using a bulk package, are per seat and difficult to transfer (MSDN subscriptions, namely). The scripts and methods used to program many *nix boxes over the past 30 years are still highly applicable to learning and building your own system. MS OTOH has published a new OS or set of tools to interact with it on average ever 2 years for the past 15. Also, the speed of MS's updates are pushed by forces that may not be required of your installation, but must be installed to patch a security risk or other problem. IE, Only by upgrading to Office 2000 do you close certain security holes, since 97 is no longer supported. You want to close the security hole, but cannot/do not want to upgrade your entire corporate environment, especially when losing backward compatability (Access 97, for example).
OVERALL The ONLY conclusion one can make about a MS to Linux TCO comparasion is by encompassing a use-case scenario driven from the most common demographics of users: corporate installations, home-users and industrial embedded applications. MS has seen their market share erode on all these fronts, due to simple arguments like the counterpoints above. No whitepaper fronted by any agency of Microsoft can refute that in the few CDs of a common Linux distro, the entire capability of their OS, their development platforms, their servers, their documentation and more (security, stability) show up for free, and work out of the box.
Real TCO in Linux can be kept to intelligence of the owners and 0 in currency. There is simply no way for Microsoft to compete against that model without offering better software. To this end, they lose when those use-case patterns do not adopt their innovations, while Linux tools methodically become more stable and approachable.
Re:Stop looking for "programming" jobs
on
Exporting Myself?
·
· Score: 1
Ah yes. Thanks, I stand corrected. (I was seeing 'A' and thinking associates). However, then, poster could certainly get a Masters, which may truely be overkill at the moment.
Re:Stop looking for "programming" jobs
on
Exporting Myself?
·
· Score: 4, Insightful
Sorry, but to be a good analyst (this IS the type of "designer" you suggest, correct?), one should put in the time, so to speak, as a programmer for awhile. Nothing is more mismatched than a book-only-learned analyst. They are good at design patterns, concepts of reuse and well-read, but give them a real-world legacy mess (or just an older interface or module) and they suffer a breakdown in productivity. "Here's fix this crappy site in 2 weeks. No you cannot rewrite it."
That "manual labor for your fingers" is actually quality time spent solving a pool of math, logic, algorithm and communication problems. Also, the trials of explaining such tasks, issues and achievements in easy-to-understand words to PHBs is valuable. Don't underestimate simply being a programmer.
To poster: I am guessing you didn't want to live in debt forever, so you skipped the BS and got just the BA. Maybe you can work part time using whatever skills relate to your field, and get the other 2 years over with. Heck, some people make a career out of attending classes just that way. As to your question, I'm at a loss. I'm doubting you'd like living as a programmer in another country...
In MY cube, I've cut the space into half-high walled partitions, each with their desk, phone and chair. They are.5 metre sq. Then when I'm sick of my surroundings, I can switch cubes! I have scenic calendars on the outer walls so I can treat myself to a "window office" now and then.
If this is "trade secret protection" request is granted, the threat of SCO suing the known Linux customer base is cracked a little wider. It means number 2 above (the code in question) can be shown in court to the offending party, but not openly.
So, yes you have to pay for the code, according to the SCOAU (SCO Alternate Universe). You could be sued, request the code in question, and still lose your case, it being sealed because of SCO's Trade Secret requests.
However, I can only see them doing it once, since even though the implementation of such a programing design could be SCO's, the location of such implementation in the Linux modules would not be. Even if it was, any next upgrade from some such entity (say SCO sues Red Hat, Red Hat settles and then their next release replaces 4 modules explicitly). We'd all know what that implies.
These kinds of books are great to inspire a population of hobbiests to write new and interesting programs (just one of a set of reader types). However, without a good grasp of the prior solutions to most technology issues, one is bound to spend a lot of time experimenting to create something that already exists.
With all the tools OS/GNU and such, there should be strong emphasis on the myriad of projects already out there. Sadly, this amount of information may be too dynamic or large for printed matter. A lot of great minds are all designing bad MP3 players, for example, when the algorithm and code is pretty much commoditized.
Eh. Don't get me wrong, I'm not to stifle innovation in existing concepts, but most subjects are vastly deeper than what a home-hobbiest is going to know when typing up their first few projects.
My phone service goes down more than my ISP. I've been able to chat to people over cable modem while my phone is dead from a storm.
If said burgler cut your phone wires, you not know it until you picked up the phone. If your IP-based phone was cut, it could let you know immediately with a type of alarm.
This is yet just another business model that needs drastic changes based on the networked digital economy. Libraries, Travel Agents, Porn, Gambling, Advertising, Music [and eventually] Movie distribution, and now telephone communication. All going to the packets.
There is something scary about so much economy on the wires, but such will be the Information Age. The telephone companies are one of the larger entities who's cheese is getting moved, and I don't expect them to go down without a crying about it. The RIAA is another who now has to accept that their content will be on the lines to stay.
Once everyone starts using end-point phones that accept not only traditional lines, but ethernet, we're going to see a very low barrier-to-entry for providing phone service, IMO. Once this steps up, all the bells and whistles we again be sold to us (photo, video, messaging, etc). There's a slew of new possibilities that a lot of new players could innavte into such a system. I'm looking forward to it.
Imagine if you will: - Scanning photos, receipts, etc into your [cell] phone for the recipient, live or to an answering machine. Or sending them a video. - Getting a message on your phone from your mother who thinks "this commercial, watch this" or "this newspaper article" was really interesting or funny. Ok, maybe this isn't all good. - Calling your home silently and hopping around the house phones to check their cameras (babysitter, teenagers, security). - Having a web page served from your phone that holds your recorded messages, images, memos, stored documents, etc. And being able to pull them out of another phone.
You use languages? Sissy! In MY DAY, we'd plug wires into a wall of vaccuum tubes. Every few hours, we'd shutdown and replace the burnouts. and don't even ASK ME about the BUGS.
I don't know anyone who wants to sit in front of their computer screen for 2 hours to watch a movie and there are very few people who have the hardware to play the movie on their tv's.
You're kidding, right? Xbox et. al. are hoping to do just this.
Most computers have DVD players. Most TV, (CRT, LCD, Plasma) have a digital in. Many Video cards have TV/digital outs.
You're right, we're not close, we've already passed this hurdle. Check out how many laptops are playing movies on the next plane ride. People who are within 2 years of their last purchase are probably only a few wires and programs away from using their computer as Tivo/DVDplr/CDplr.
So assuming hashes are 0.001% of the text, I come up with 100 days. Reading that the tables are kept, we're talking about 100 days to form a full hash table and crypt() is dead ?
PS no offense taken. My crypto skills are in the basement, but I enjoy learning. I think people wanted the comment to mod up to get answered, that's all.
Long filename extensions to the 8.3 format in FAT is the detail here. It applies to devices that come formatted to read/write this extension.
The details to the 8.3 format extensions for long filenames involve using a previously unused portion of the filename (plus some of what's used, see the tilde) and then expanding the table such filenames are stored to form name chains.
But there are several less expensive ways to create such a system. We've talked before about metadata. Although I'm a fan of embedding metadata into the file, the coupling should be weak. Nothing says a camera (for example) cannot keep an XML-format metadata in another file, linked to an 8.3 filename (using all 255 char values in the 8.3). Upon extraction to the external ports, new filenames can be created from the metadata (say, the NAME element) and replace the 8.3 key.
Or, simply export the 8.3 and the metadata in file pairs to the external device. Users can sort, annotate their media (even with voice, as my camera does, which I never use).
I hand MS credit for innovating this solution. I don't mind them cashing in. However, the smart silient majority should start thinking up a new bunch of ideas to eventually become a standard candidate.
I see this as a single-minded business. SPAMcop wants to remove spammers by hunting down the true origins of mail it is told are illegitimate, or through filters. OTOH, it is owned by a company that teaches and sells mass-marketing schemes. This mass-marketer has competition, and thats exactly what SPAMcop will be going after. Bingo! You have a great model to improve your scores by showing actual tallies of improved responses from people using your lists or methods.
However, there will be an easy way to detect this: If the companies that are sending the spam are ignored by SPAMcop and also part of their enterprise, we have the feared result. At the moment, I haven't seen any evidence of this posted anywhere. But I'm only one person.
mug
Re:where are the open source XML repositories
on
Effective XML
·
· Score: 1
no idea? think about it guy.
An open DTD repository would serve no purpose. For example: A Photo. ANY definition of a "standard" for Name, Location, Date, Photographer and Content in a DTD repository would immediately fragment. Further tags would be defined (film used, camera, settings, ambient light levels, digital source information) or other esoteric tags for specialized uses.
So then, we start another DTD, layered on top of/extending the prior (or not, where unknowingly someone starts a new one) for the professional printer or photographer, etc. Of course, then we get into content tags (PR0N based, geographic, designer-esque terms). It's the long slow trail to a failed Grand-Unification Theory of information. It takes a dictatorship to run, which clears up some mess while now being an entity to appeal and fight with about standards.
You say yourself, CONTEXT is king. I agree, and so the components of metadata (ignore the XML strawman for a second) are context sensitive. There are any many contexts as uses. Imposing a fixed list (or tree) of contexts on information has this fuzzy appealing concept of "global search and find, a univeral catelog" - but it's a red herring. The structure of metadata gets knocked about in every instance until it appears as a all-in-one gigantic structure (EDI comes to mind) or woefully inadequate.
- Search your favorite P2P, find a universal tagging scheme? Even with the MP3 v1 and v2 embedded tags, it's a mess. - How many extensions of C are there? - We've seen photo album fomats come with every package. Why didn't anyone reuse an existing format? Because they thought it inadequate.
Now getting into XML, building a database from such a source has been verified over and over: bad idea. Parsing, cleanup, indexing is all great for a one-time hit, but not if XML is the primary data store. XML is the usage of an arbitrary grammar. It follows the same rules of context and reuse than everything has, from programming languages to other structured metadata : good only to a point, and always behind the times at the general level.
"To reuse the tools from another vendor" is a great goal, but each vendor may *not* want to be compatible, for many reasons. Also, if they can provide a catelog of *their* images with extensive behavior, why change? a slew of big vendors (AutoDesk, Adobe, Microsoft, Bently) are not cross-compatable because they don't want you to change. Read other formats, but do not write.
CLOSED DTD repositories work great. GM has a large library of DTDs that its vendors must obey. But there is no central repository because there is no central "context owner". Much like a linguistics issue, in my mind.
mug
Re:where are the open source XML repositories
on
Effective XML
·
· Score: 0, Interesting
No. Because XML is supposed to be extensible. The closest thing may be a XLST that taught everyone looking at your file what tags you had to display and their data types. But as for standardizing - everything that gets fixed becomes a limit to be broken some day.
Other than that, you're defining a file format. Why bother? Many formats already exist that encapsulate some form of metadata. Keeping the metadata with the file is another whole subject anyway. Metadata itself should be connect-able to the target file, but not bound. There's a lot of thought about this already going around.
I can agree with you, and have tried to do similar with my family at time. However, there is a certain level of accomodation you must have for them when they've lived an entire life equating "giving a gift" with "expressing thanks/love/etc" during the holiday season. When I've run into these folks, I try to steeer them towards more and more practical ideas for gifts (mostly outdoorsy-type stuff since I like hiking). There is a limit for everything though, where you feel like you have "so much stuff" that getting more gifts defeats the purpose.
This year, I've convinced people to make a donation to a charity of their choice in my name, then mail me a card or glossy about the charity as the gift. Seems like people are really into it too.
What happens when SCO's legal bills or the SCOX price drop so that the lawfirm pretty much owns most of ther floating stock? At that point, wouldn't IBM be fighting a legal firm more than any real tech company?
Boise, formerly SCO, Inc, is now taking up the legal fight to collect damages from the defunct product of SCO's. After firing the last of the technical personnel from SCO, a spokesman for Boise said "We really don't need any technical information any more. As long as we can push out press releases and manage the SCOX stock price."
Wuld you write an entire [insert utility of choice] or use the one that came with the OS? When I use windows, part of knowing how to interact with the machine is adapting to it's special requirements, like a GUI. When I use a Linux, I again adapt and learn the Command Line Utilities (CMU) and config files that must be tailored to suit my needs.
But either way, I'm not spending every waking hour versing myself on each tool. It's only when I need them that I delve into them. If someone asks for my help, I can either say "yes, I've worked with that before" or "no, that may take some time to learn". The time it takes is relative to your experience.
And that brings me to my point, CompSci schooling (or just plain 'ole experience) teaches one not to fear the machine or be overwhelemed by it's tools. You can approach a box and start playing around with it, knowing that within certain bounds, you're going to learning things. Try the command line switches, help files, man pages, internet sites for ramping up. Run it 3 ways to sunday to see how it reacts.
As complex components become commoditized, people are learning CMU interaction with the machine where the block boxes are larger and larger. Thats fine. Even if they don't understand those low-level details (which now can ecome trivia but not essential for most jobs), they are productive. So you have to rank yourself on end-user productivity (knowing the blocks ahead of time, even that they exist) or technical prowess (know how to fix/replace any particular block). I think as programmers we all want to be both, but compsci is just too big nowadays.
I think they're offering discounts, not just paying people. Of course, I may be reading this wrongly, but there will be no checks written and accompanying a letter that says "here's our thanks for switching to FreeBSD!"
Slashdot Mirror
I'm going to take the Stallman position here. People SHOULD build, perfect and extend these things because - eventually - they will contribute to keeping information free.
;)
I agree with you about consumer viability for these PVR applications. But your position can be used to dissuade home-built versions of a lot of things. Getting them off the ground is part of the fun, insanity and adventure of geekdom. There ARE success stories in that realm.
Now, I don't own a TV, so I come already doubting the value of such information, but that's another story. After all your "saving of money" you're going to watch what drivel?
Counterpoint: The methods for administering Linux boxen must be measured in several ways:
Cost to install and initially setup OS for intergration into an environment, including security
Cost build and configure a default user account
Cost to write/configure health monitoring scripts for indication of adverse events
Backup/Restore capabilities and costs
.NET development, SQLServer, IIS, Exchange servers, etc all come with an ongoing license fee (License 6 model). Also comparing ramp-up time for each of these tools may involve training materials which MS does not give away, versus Linux information flow online and in published info/man pages the costs are stacked highly against Microsoft.
This is only to scratch the surface of admin. In detail, a Linux admin has many capabilities transferrable from older Unix-style OS's. Windows OS's, OTOH, have routinely changed their server administration methods, both through updates and through versions. NT/2000/XP each have delicate nuiances that must be published. This doesn't mention services like ActiveDirectory etc that can radically change a administration model.
More efficiency per dollar
Counterpoint: Not counting continuous uptime, Microsoft and Linux can each be tuned to comparable speeds for dedicated servers (mail, file, print, web). This immediately clouds when comparing details since scripting hosts, programming style, cache arrangements and developer expertise may be vastly different. However, given that most tools for Linux are also free, Microsoft must admit the costs for not only the OS, but all their server offerings.
NOTE: Counting continuous uptime, MS loses by such a overwhelming amount that it makes the above counterpoint moot. Cycling a Microsoft box because of a update (forced or not) - across all installed packages - is not uncommon. Keeping the Microsoft tool compatability across these installations is also a hit or miss operation with a poor historical record.
Reduced development costs
Counterpoint: Installing all tools from a end-user Linux distro is under the "free" cost of the OS itself. The tools from Microsoft, best bought in a corporate level using a bulk package, are per seat and difficult to transfer (MSDN subscriptions, namely). The scripts and methods used to program many *nix boxes over the past 30 years are still highly applicable to learning and building your own system. MS OTOH has published a new OS or set of tools to interact with it on average ever 2 years for the past 15. Also, the speed of MS's updates are pushed by forces that may not be required of your installation, but must be installed to patch a security risk or other problem. IE, Only by upgrading to Office 2000 do you close certain security holes, since 97 is no longer supported. You want to close the security hole, but cannot/do not want to upgrade your entire corporate environment, especially when losing backward compatability (Access 97, for example).
OVERALL
The ONLY conclusion one can make about a MS to Linux TCO comparasion is by encompassing a use-case scenario driven from the most common demographics of users: corporate installations, home-users and industrial embedded applications. MS has seen their market share erode on all these fronts, due to simple arguments like the counterpoints above. No whitepaper fronted by any agency of Microsoft can refute that in the few CDs of a common Linux distro, the entire capability of their OS, their development platforms, their servers, their documentation and more (security, stability) show up for free, and work out of the box.
Real TCO in Linux can be kept to intelligence of the owners and 0 in currency. There is simply no way for Microsoft to compete against that model without offering better software. To this end, they lose when those use-case patterns do not adopt their innovations, while Linux tools methodically become more stable and approachable.
Ah yes. Thanks, I stand corrected. (I was seeing 'A' and thinking associates). However, then, poster could certainly get a Masters, which may truely be overkill at the moment.
Sorry, but to be a good analyst (this IS the type of "designer" you suggest, correct?), one should put in the time, so to speak, as a programmer for awhile. Nothing is more mismatched than a book-only-learned analyst. They are good at design patterns, concepts of reuse and well-read, but give them a real-world legacy mess (or just an older interface or module) and they suffer a breakdown in productivity. "Here's fix this crappy site in 2 weeks. No you cannot rewrite it."
That "manual labor for your fingers" is actually quality time spent solving a pool of math, logic, algorithm and communication problems. Also, the trials of explaining such tasks, issues and achievements in easy-to-understand words to PHBs is valuable. Don't underestimate simply being a programmer.
To poster: I am guessing you didn't want to live in debt forever, so you skipped the BS and got just the BA. Maybe you can work part time using whatever skills relate to your field, and get the other 2 years over with. Heck, some people make a career out of attending classes just that way. As to your question, I'm at a loss. I'm doubting you'd like living as a programmer in another country...
Actually, the SCO drama only really affects investors, and has little to do with technology at this point.
In MY cube, I've cut the space into half-high walled partitions, each with their desk, phone and chair. They are .5 metre sq. Then when I'm sick of my surroundings, I can switch cubes! I have scenic calendars on the outer walls so I can treat myself to a "window office" now and then.
No, don't do it man! Think of the poor submarines! "Stranded!"
Actually, parent said:
You are kinda off-base in your criticism. Attitude aside, parent had a good point.
If this is "trade secret protection" request is granted, the threat of SCO suing the known Linux customer base is cracked a little wider. It means number 2 above (the code in question) can be shown in court to the offending party, but not openly.
So, yes you have to pay for the code, according to the SCOAU (SCO Alternate Universe). You could be sued, request the code in question, and still lose your case, it being sealed because of SCO's Trade Secret requests.
However, I can only see them doing it once, since even though the implementation of such a programing design could be SCO's, the location of such implementation in the Linux modules would not be. Even if it was, any next upgrade from some such entity (say SCO sues Red Hat, Red Hat settles and then their next release replaces 4 modules explicitly). We'd all know what that implies.
I know we all hate and despise spam, but doesn't a possible 20 years in the clink sound a bit excessive?
Howabout 20 years clicking away on unwanted popups?
These kinds of books are great to inspire a population of hobbiests to write new and interesting programs (just one of a set of reader types). However, without a good grasp of the prior solutions to most technology issues, one is bound to spend a lot of time experimenting to create something that already exists.
With all the tools OS/GNU and such, there should be strong emphasis on the myriad of projects already out there. Sadly, this amount of information may be too dynamic or large for printed matter. A lot of great minds are all designing bad MP3 players, for example, when the algorithm and code is pretty much commoditized.
Eh. Don't get me wrong, I'm not to stifle innovation in existing concepts, but most subjects are vastly deeper than what a home-hobbiest is going to know when typing up their first few projects.
My phone service goes down more than my ISP. I've been able to chat to people over cable modem while my phone is dead from a storm.
If said burgler cut your phone wires, you not know it until you picked up the phone. If your IP-based phone was cut, it could let you know immediately with a type of alarm.
This is yet just another business model that needs drastic changes based on the networked digital economy. Libraries, Travel Agents, Porn, Gambling, Advertising, Music [and eventually] Movie distribution, and now telephone communication. All going to the packets.
There is something scary about so much economy on the wires, but such will be the Information Age. The telephone companies are one of the larger entities who's cheese is getting moved, and I don't expect them to go down without a crying about it. The RIAA is another who now has to accept that their content will be on the lines to stay.
Once everyone starts using end-point phones that accept not only traditional lines, but ethernet, we're going to see a very low barrier-to-entry for providing phone service, IMO. Once this steps up, all the bells and whistles we again be sold to us (photo, video, messaging, etc). There's a slew of new possibilities that a lot of new players could innavte into such a system. I'm looking forward to it.
Imagine if you will:
- Scanning photos, receipts, etc into your [cell] phone for the recipient, live or to an answering machine. Or sending them a video.
- Getting a message on your phone from your mother who thinks "this commercial, watch this" or "this newspaper article" was really interesting or funny. Ok, maybe this isn't all good.
- Calling your home silently and hopping around the house phones to check their cameras (babysitter, teenagers, security).
- Having a web page served from your phone that holds your recorded messages, images, memos, stored documents, etc. And being able to pull them out of another phone.
mug
You use languages? Sissy! In MY DAY, we'd plug wires into a wall of vaccuum tubes. Every few hours, we'd shutdown and replace the burnouts. and don't even ASK ME about the BUGS.
Languages are just portals for virii!
I don't know anyone who wants to sit in front of their computer screen for 2 hours to watch a movie and there are very few people who have the hardware to play the movie on their tv's.
You're kidding, right? Xbox et. al. are hoping to do just this.
Most computers have DVD players.
Most TV, (CRT, LCD, Plasma) have a digital in.
Many Video cards have TV/digital outs.
You're right, we're not close, we've already passed this hurdle. Check out how many laptops are playing movies on the next plane ride. People who are within 2 years of their last purchase are probably only a few wires and programs away from using their computer as Tivo/DVDplr/CDplr.
So assuming hashes are 0.001% of the text, I come up with 100 days. Reading that the tables are kept, we're talking about 100 days to form a full hash table and crypt() is dead ?
PS no offense taken. My crypto skills are in the basement, but I enjoy learning. I think people wanted the comment to mod up to get answered, that's all.
Even so, using a 10 character input of about 72 possible input chars, isn't 207 billion still only like .0000055% of the total search space?
So that 20000 * 80minutes gives ~1% of the space cracked?
2000000 * 80 minutes = 304 years to fully close the space.
With a perfect distribution, the mean of about 150 years seems like a long time.
Someone please check my assumptions here.
Long filename extensions to the 8.3 format in FAT is the detail here. It applies to devices that come formatted to read/write this extension.
The details to the 8.3 format extensions for long filenames involve using a previously unused portion of the filename (plus some of what's used, see the tilde) and then expanding the table such filenames are stored to form name chains.
But there are several less expensive ways to create such a system. We've talked before about metadata. Although I'm a fan of embedding metadata into the file, the coupling should be weak. Nothing says a camera (for example) cannot keep an XML-format metadata in another file, linked to an 8.3 filename (using all 255 char values in the 8.3). Upon extraction to the external ports, new filenames can be created from the metadata (say, the NAME element) and replace the 8.3 key.
Or, simply export the 8.3 and the metadata in file pairs to the external device. Users can sort, annotate their media (even with voice, as my camera does, which I never use).
I hand MS credit for innovating this solution. I don't mind them cashing in. However, the smart silient majority should start thinking up a new bunch of ideas to eventually become a standard candidate.
mug
I see this as a single-minded business. SPAMcop wants to remove spammers by hunting down the true origins of mail it is told are illegitimate, or through filters. OTOH, it is owned by a company that teaches and sells mass-marketing schemes. This mass-marketer has competition, and thats exactly what SPAMcop will be going after. Bingo! You have a great model to improve your scores by showing actual tallies of improved responses from people using your lists or methods.
However, there will be an easy way to detect this: If the companies that are sending the spam are ignored by SPAMcop and also part of their enterprise, we have the feared result. At the moment, I haven't seen any evidence of this posted anywhere. But I'm only one person.
mug
no idea? think about it guy.
An open DTD repository would serve no purpose. For example: A Photo. ANY definition of a "standard" for Name, Location, Date, Photographer and Content in a DTD repository would immediately fragment. Further tags would be defined (film used, camera, settings, ambient light levels, digital source information) or other esoteric tags for specialized uses.
So then, we start another DTD, layered on top of/extending the prior (or not, where unknowingly someone starts a new one) for the professional printer or photographer, etc. Of course, then we get into content tags (PR0N based, geographic, designer-esque terms). It's the long slow trail to a failed Grand-Unification Theory of information. It takes a dictatorship to run, which clears up some mess while now being an entity to appeal and fight with about standards.
You say yourself, CONTEXT is king. I agree, and so the components of metadata (ignore the XML strawman for a second) are context sensitive. There are any many contexts as uses. Imposing a fixed list (or tree) of contexts on information has this fuzzy appealing concept of "global search and find, a univeral catelog" - but it's a red herring. The structure of metadata gets knocked about in every instance until it appears as a all-in-one gigantic structure (EDI comes to mind) or woefully inadequate.
- Search your favorite P2P, find a universal tagging scheme? Even with the MP3 v1 and v2 embedded tags, it's a mess.
- How many extensions of C are there?
- We've seen photo album fomats come with every package. Why didn't anyone reuse an existing format? Because they thought it inadequate.
Now getting into XML, building a database from such a source has been verified over and over: bad idea. Parsing, cleanup, indexing is all great for a one-time hit, but not if XML is the primary data store. XML is the usage of an arbitrary grammar. It follows the same rules of context and reuse than everything has, from programming languages to other structured metadata : good only to a point, and always behind the times at the general level.
"To reuse the tools from another vendor" is a great goal, but each vendor may *not* want to be compatible, for many reasons. Also, if they can provide a catelog of *their* images with extensive behavior, why change? a slew of big vendors (AutoDesk, Adobe, Microsoft, Bently) are not cross-compatable because they don't want you to change. Read other formats, but do not write.
CLOSED DTD repositories work great. GM has a large library of DTDs that its vendors must obey. But there is no central repository because there is no central "context owner". Much like a linguistics issue, in my mind.
mug
No. Because XML is supposed to be extensible. The closest thing may be a XLST that taught everyone looking at your file what tags you had to display and their data types. But as for standardizing - everything that gets fixed becomes a limit to be broken some day.
Other than that, you're defining a file format. Why bother? Many formats already exist that encapsulate some form of metadata. Keeping the metadata with the file is another whole subject anyway. Metadata itself should be connect-able to the target file, but not bound. There's a lot of thought about this already going around.
mug
I can agree with you, and have tried to do similar with my family at time. However, there is a certain level of accomodation you must have for them when they've lived an entire life equating "giving a gift" with "expressing thanks/love/etc" during the holiday season. When I've run into these folks, I try to steeer them towards more and more practical ideas for gifts (mostly outdoorsy-type stuff since I like hiking). There is a limit for everything though, where you feel like you have "so much stuff" that getting more gifts defeats the purpose.
This year, I've convinced people to make a donation to a charity of their choice in my name, then mail me a card or glossy about the charity as the gift. Seems like people are really into it too.
mug
What happens when SCO's legal bills or the SCOX price drop so that the lawfirm pretty much owns most of ther floating stock? At that point, wouldn't IBM be fighting a legal firm more than any real tech company?
Boise, formerly SCO, Inc, is now taking up the legal fight to collect damages from the defunct product of SCO's. After firing the last of the technical personnel from SCO, a spokesman for Boise said "We really don't need any technical information any more. As long as we can push out press releases and manage the SCOX stock price."
Wuld you write an entire [insert utility of choice] or use the one that came with the OS? When I use windows, part of knowing how to interact with the machine is adapting to it's special requirements, like a GUI. When I use a Linux, I again adapt and learn the Command Line Utilities (CMU) and config files that must be tailored to suit my needs.
But either way, I'm not spending every waking hour versing myself on each tool. It's only when I need them that I delve into them. If someone asks for my help, I can either say "yes, I've worked with that before" or "no, that may take some time to learn". The time it takes is relative to your experience.
And that brings me to my point, CompSci schooling (or just plain 'ole experience) teaches one not to fear the machine or be overwhelemed by it's tools. You can approach a box and start playing around with it, knowing that within certain bounds, you're going to learning things. Try the command line switches, help files, man pages, internet sites for ramping up. Run it 3 ways to sunday to see how it reacts.
As complex components become commoditized, people are learning CMU interaction with the machine where the block boxes are larger and larger. Thats fine. Even if they don't understand those low-level details (which now can ecome trivia but not essential for most jobs), they are productive. So you have to rank yourself on end-user productivity (knowing the blocks ahead of time, even that they exist) or technical prowess (know how to fix/replace any particular block). I think as programmers we all want to be both, but compsci is just too big nowadays.
I think they're offering discounts, not just paying people. Of course, I may be reading this wrongly, but there will be no checks written and accompanying a letter that says "here's our thanks for switching to FreeBSD!"