... as usual, with MS "idea(l)s, the fault will be in the implementation.
"Say a consumer sees an ad for a product in a video ad one day, and then clicks on a text ad to visit the retailer's site the next day, and then eventually sees a banner ad that leads to a purchase. All of the monetary credit tends to go to the text link that was clicked on."
Okay, I agree. It is a proven fact that the most effective advertising is based on 'reminders', and true that it is only the last 'sale tripper' ad that gets all the credit - and revenue. Yes, the system needs to be reworked, especially for pay-by-click-only ads.
Take ads such as Diesel's: It's all in the branding, and they couldn't give a flying (expletive) if you visit their website; clients who do so only represent an infitismal persentage of their client base.
So, how to calculate "what and who makes a sale", and how to distribute the revenues fairly? Who placed the ad most prominently on their webpage (and to what effect), or who rigged their site to display the add next to relevent content (and to what effect?)? How far down the "ad chain" was that last click (and how are you going to calculate that - track and/or hunt down every movement of the person who clicked on an ad?)? They would have to build a database that logs the IP of each and every person who downloads an ad participating in their "system"`; I don't welcome the idea.
Page rank is a gauge of popularity, not content, more than anything. It's a factor that only comes into play when google's algorithm judges your content at the same level of that of another page as an answer to a query - only then the most popular page gets top spot.
I like the concept of a semantic web, but frankly, I don't like its present trend of implication. It seems so "chunky" (metacruft), and still has to be managed by humans if it hopes to attain any level of accuracy.
If we can't mimic human thought, perhaps we can make a search method that can take into account the results of its reasoning. Boolean searches are quite powerful - why not develop a system along those lines? With added functionality - say, "bob" -5 "Ralph" would turn up pages that have those two words within five words of each other, with results ordered by relevance (matching boolean '-5', matching boolean 'AND', matching one word, etc.). Have the boolean markup generated by a GUI, if you will. As for prices, I'm sure these could be recognised by any search engine if it is programmed correctly.
Even this solution does not seem "complete" to me - somehow we're going to have to find a way to recreate human reasoning (to a certain level) before we can have a semantic web that is of any widespread (www) use to anyone.
Hmmm, perhaps it won't be so hard to police - block everything but port 80, 109, 110 and 366 for your virtual IP (web, pop, pop3 and SMTP respectively), constrain ftp access to only the site's root folder (and upwards), and limit database queries to localhost-only. I'm sure there's ways around that (and sorry, Plesk isn't that), but high bw traffic is always detectable.
As for parent, my boss said something in the same regard only a couple days ago - we are starting a secure online photo archival service - when he saw a similar deal. He suggested we sign up and upload a few TB just to see how far "unlimited" goes : )
I really didn't understand how they can offer "unlimited" disk space until I read this in their "terms and conditions" (the fine print, whot):
"You acknowledge that the Web Hosting service is offered as a platform to host and serve web pages and web sites and is not offered for other purposes, such as remote disk space storage. Accordingly, You understand and agree to use the web hosting service solely for the purpose of hosting and serving web pages as viewed through a web browser and the Hypertext Markup Language (HTML) or other equivalent technology."
So in other words, websites and databases (for websites) only (will they police this?). Yet there are companies out there offering unlimited disk space for remote storage, and how they go about this, I have yet to understand. Perhaps something along the same lines as Yahoo's former offering?
If you're concerned about more than "sending things home", then I suggest getting a VAIO. It pains me to say this as a mac-accolyte, but as far as hardware is concerned, I've seen nothing more portable nor light in the way of fully-operational. Hell, buy one and install linux on it - a bit more configuration before your trip, but you can be sure that it works everywhere.
Both the above and grandparent have a point. Remote management GUI's are potential targets - but users need GUI's.
Administration work should really most safely be done through a monitor and keyboard plugged directly into the server, and any gui-like access/interface should only exist on the administrator's computer. Yet would it still be safe to manage a computer ~remotely~ if the gui existed only on the administrator computer?
This would mean setting up ~minimal~ GUI's for those web-users - meaning mail access, (sub)domain management, etc. - a lot of work for someone.
Sure, they help people to help themselves, as long as they help themselves - their way. Of course those self-evident "teachings" gives them license (they think) to help themselves to their earnings.
Yep, if they made hackey code, they'd better fix it. If MS really was forward-thinking, they'd target their meta-tag idea to the same, not newly-created pages. The idea of someone working to make a page standards-compliant, testing that page through W3 to prove the same, then having to add a MS tag to each and every page - just for the sake of ONE browser and version - just to state again the same is backwards and ridiculous.
Hear hear. MS should know by now the hacks WM's use to correct its failings, so instead of shipping with three rendering engines, why not ship with one plus a "filter" that could parse bad (IE-targeted) code and render it correctly? Better still, why not tag pages coded for OLD versions of IE, and have the new version render correctly by default?
No, it won't - and it is the test that sets its own rules and conditions, not MS.
The Acid2 test is for any and every browser. Bare-bones code, no cheating allowed. Or do you really think that the Acid2 test page will stick MS's meta tag into its header? Not. So unless it renders correctly by default, IE8 will fail. Period.
But you can't compare all OS's and applications as if they all have the same work ethics, ideas and development resources.
MS is a company that has had a) few original ideas but b) HUGE resources - the difference between the two, in addition to their decades of constant, stubborn fuck-ups, smacks of mismanagement and sheer technological incompetence, making them one of the "worst" (inexcusable) software companies out there; or perhaps they simply don't care, as, in spite of all their product's faults, people keep buying them.
...and that extra meta tag is just a peevy way of trying to oblige everyone to think "Microsoft! Microsoft! (is everywhere!)". MS and Karl Rove must share the same condescending views on humanity - and of course, even with all their respective faults and failings, think themselves above it - and spend most of (their|other people's) time and money trying to (convince themselves|convey to us) that they weren't at fault - without really saying it. Take this phrase for example:
"IE6's rendering behavior was not updated for five years, leading many developers to assume its rendering was both accurate and unlikely to change." ..."Accurate"!? Not a word I would use to describe that browser's rendering - nor one any decent webmaster would use to describe the same, either. What a load of two-bit high-school lawyer spin - that is supposed to "cover up" six years of complete - and stubborn - failure?
I was initially happy that MS wised up and decided to play straight with their browser - happy FOR them - but now I couldn't give a fuck. Talk about a company wasting a chance at recieving a widespread feeling of goodwill from the thinking community - the first since decades, if ever.
Err, fibre. Yes, but Expensive. Who provides your interconnectivity? If you have a reliable ISP (or better still, a TC company such as Colt) who can rent you a U or two for your central server AND fibre-connect you to it, you should have little to no problems - with added security. I'm sure that for an added charge they will (or you can) set up automated backups for you, or you can implement something like a mirrored RAID for even more security - but that depends on your needs.
I would also suggest dividing server roles between "tasks" and "data". IMHO, for many tasks, such as printing, a server is not even needed - a networked machine will suffice. Data should be secure and always accessible from all points; tasks, as they depend on the machines they control, should be distributed to points closest to them. If the printer's LAN loses web access, its print server, if it is in another remote location, becomes a... print bank. A role a bit much for the investment in an entire server/setup.
The question is so broad, it's hard to answer. The definition of "good" and "cheap" differ for most everyone in this concern.
I've recently had to set up such a system, and I opted for... VPN. Secure (because I opted for the ssh certificates version) for sure, but can be slow as molasses for the (uploading) remote connection, depending on their bandwidth. Yet both central office and main outposts have (the European equivalent of) T1 connections, with a secondary backup connection option if needed. DynDns is a great option for nailing down those company IP's through all locales and differing ISP behaviours.
Whether each interconnected LAN has its own server has really nothing to do with interconnectivity. Here all "remote" connections take/depose their data from our central office server (Windows 2003 - aaaargh), but in the local LAN the server is not really a server at all, rather a backup utility. I might add that I'll be reworking the above soon into a system directly connected to our (new) provider through fibre, but this changes nothing in the workings of the point-to-point connections. Build your system architecture only according to your (technological) needs.
Print queues should only be run by a machine directly linked to the printer that will print them -- and you don't necessarily need a server to fulfil that task. Unless, of course, you're a printer.
Slashdot Mirror
I'm an admitted Mac fanboy, but it would be really helpful for other readers if I could 'Wiki' you - can you cite your sources? I could use them, too.
How about some revised parsing for those front-page titles, webmaster?
...would that be the sound of them falling?
Take ads such as Diesel's: It's all in the branding, and they couldn't give a flying (expletive) if you visit their website; clients who do so only represent an infitismal persentage of their client base.
So, how to calculate "what and who makes a sale", and how to distribute the revenues fairly? Who placed the ad most prominently on their webpage (and to what effect), or who rigged their site to display the add next to relevent content (and to what effect?)? How far down the "ad chain" was that last click (and how are you going to calculate that - track and/or hunt down every movement of the person who clicked on an ad?)? They would have to build a database that logs the IP of each and every person who downloads an ad participating in their "system"`; I don't welcome the idea.
...not to mention "Divorce court".
Oh, wait, this is slashdot...
Parent has actually provided the solution. Forget Steve Jobs, someone send that guy to /... I hope he reads the whole thread.
Page rank is a gauge of popularity, not content, more than anything. It's a factor that only comes into play when google's algorithm judges your content at the same level of that of another page as an answer to a query - only then the most popular page gets top spot.
I like the concept of a semantic web, but frankly, I don't like its present trend of implication. It seems so "chunky" (metacruft), and still has to be managed by humans if it hopes to attain any level of accuracy.
If we can't mimic human thought, perhaps we can make a search method that can take into account the results of its reasoning. Boolean searches are quite powerful - why not develop a system along those lines? With added functionality - say, "bob" -5 "Ralph" would turn up pages that have those two words within five words of each other, with results ordered by relevance (matching boolean '-5', matching boolean 'AND', matching one word, etc.). Have the boolean markup generated by a GUI, if you will. As for prices, I'm sure these could be recognised by any search engine if it is programmed correctly.
Even this solution does not seem "complete" to me - somehow we're going to have to find a way to recreate human reasoning (to a certain level) before we can have a semantic web that is of any widespread (www) use to anyone.
Um. Oops. Answer to question... my own?
Hmmm, perhaps it won't be so hard to police - block everything but port 80, 109, 110 and 366 for your virtual IP (web, pop, pop3 and SMTP respectively), constrain ftp access to only the site's root folder (and upwards), and limit database queries to localhost-only. I'm sure there's ways around that (and sorry, Plesk isn't that), but high bw traffic is always detectable.
As for parent, my boss said something in the same regard only a couple days ago - we are starting a secure online photo archival service - when he saw a similar deal. He suggested we sign up and upload a few TB just to see how far "unlimited" goes : )
So in other words, websites and databases (for websites) only (will they police this?). Yet there are companies out there offering unlimited disk space for remote storage, and how they go about this, I have yet to understand. Perhaps something along the same lines as Yahoo's former offering?
Okay, mod me -1 for spelling and +1 for posting so late : )
If you're concerned about more than "sending things home", then I suggest getting a VAIO. It pains me to say this as a mac-accolyte, but as far as hardware is concerned, I've seen nothing more portable nor light in the way of fully-operational. Hell, buy one and install linux on it - a bit more configuration before your trip, but you can be sure that it works everywhere.
Mom was probably trained on Windows, and didn't get that a) everything is in one place b) in about half the clicks away.
Yes, but not only do they not have a policy: they can't define what "victory" is, either.
Uh, the "freedom" part. Freedom to be wiretapped and misinformed?
Basically the Bush "internets plan" paper was "Yay! We did nothing - what problems? - and we like it that way!"
Both the above and grandparent have a point. Remote management GUI's are potential targets - but users need GUI's.
Administration work should really most safely be done through a monitor and keyboard plugged directly into the server, and any gui-like access/interface should only exist on the administrator's computer. Yet would it still be safe to manage a computer ~remotely~ if the gui existed only on the administrator computer?
This would mean setting up ~minimal~ GUI's for those web-users - meaning mail access, (sub)domain management, etc. - a lot of work for someone.
Sure, they help people to help themselves, as long as they help themselves - their way. Of course those self-evident "teachings" gives them license (they think) to help themselves to their earnings.
Yep, if they made hackey code, they'd better fix it. If MS really was forward-thinking, they'd target their meta-tag idea to the same, not newly-created pages. The idea of someone working to make a page standards-compliant, testing that page through W3 to prove the same, then having to add a MS tag to each and every page - just for the sake of ONE browser and version - just to state again the same is backwards and ridiculous.
Hear hear. MS should know by now the hacks WM's use to correct its failings, so instead of shipping with three rendering engines, why not ship with one plus a "filter" that could parse bad (IE-targeted) code and render it correctly? Better still, why not tag pages coded for OLD versions of IE, and have the new version render correctly by default?
No, it won't - and it is the test that sets its own rules and conditions, not MS.
The Acid2 test is for any and every browser. Bare-bones code, no cheating allowed. Or do you really think that the Acid2 test page will stick MS's meta tag into its header? Not. So unless it renders correctly by default, IE8 will fail. Period.
But you can't compare all OS's and applications as if they all have the same work ethics, ideas and development resources.
MS is a company that has had a) few original ideas but b) HUGE resources - the difference between the two, in addition to their decades of constant, stubborn fuck-ups, smacks of mismanagement and sheer technological incompetence, making them one of the "worst" (inexcusable) software companies out there; or perhaps they simply don't care, as, in spite of all their product's faults, people keep buying them.
I was initially happy that MS wised up and decided to play straight with their browser - happy FOR them - but now I couldn't give a fuck. Talk about a company wasting a chance at recieving a widespread feeling of goodwill from the thinking community - the first since decades, if ever.
Err, fibre. Yes, but Expensive. Who provides your interconnectivity? If you have a reliable ISP (or better still, a TC company such as Colt) who can rent you a U or two for your central server AND fibre-connect you to it, you should have little to no problems - with added security. I'm sure that for an added charge they will (or you can) set up automated backups for you, or you can implement something like a mirrored RAID for even more security - but that depends on your needs.
I would also suggest dividing server roles between "tasks" and "data". IMHO, for many tasks, such as printing, a server is not even needed - a networked machine will suffice. Data should be secure and always accessible from all points; tasks, as they depend on the machines they control, should be distributed to points closest to them. If the printer's LAN loses web access, its print server, if it is in another remote location, becomes a... print bank. A role a bit much for the investment in an entire server/setup.
The question is so broad, it's hard to answer. The definition of "good" and "cheap" differ for most everyone in this concern.
I've recently had to set up such a system, and I opted for... VPN. Secure (because I opted for the ssh certificates version) for sure, but can be slow as molasses for the (uploading) remote connection, depending on their bandwidth. Yet both central office and main outposts have (the European equivalent of) T1 connections, with a secondary backup connection option if needed. DynDns is a great option for nailing down those company IP's through all locales and differing ISP behaviours.
Whether each interconnected LAN has its own server has really nothing to do with interconnectivity. Here all "remote" connections take/depose their data from our central office server (Windows 2003 - aaaargh), but in the local LAN the server is not really a server at all, rather a backup utility. I might add that I'll be reworking the above soon into a system directly connected to our (new) provider through fibre, but this changes nothing in the workings of the point-to-point connections. Build your system architecture only according to your (technological) needs.
Print queues should only be run by a machine directly linked to the printer that will print them -- and you don't necessarily need a server to fulfil that task. Unless, of course, you're a printer.