The internet will continue to grow in capacity and as it has new products will come out to fill the void.
My biggest issues with youtube are at work in our main office. We have a large application hosted in data center. It is a major hub for internet connectivity for the region. Given that we are so close to some big vendors we can get lots of bandwith for relatively low prices. If my employees where sitting in that facility they could surf youtube.com all day.
Now at home I can also do it. I pay Comcast a big more for the extra bandwidth and I can download over a meg a second from some sites. Verizon is going to be laying fiber directly to houses and businesses soon.
Get into our offices and it is a different story. We have dual t1s coming in and only 60+ employees, but we are constantly saturated. Combine that with the fact that Cisco Pixes have horrible throttling support and you end up with times when I can't even access basic websites very quickly. The issue here is that T1s and DS3s are freakin expensive compared to a simple cable modem. We have been tempted to get Comcast bussines ( which makes me shiver a bit ) because I can get larger down pipes for general internet surfing. We only host a few services such as email here so it isn't like we need megs of up bandwith.
Throttling would go along way to solve this issue. Youtube could buffer people down quite a bit, you would just have to wait for the movie to buffer a bit. For shared internet connections and ISPs this could allow for better QOS.
Distribution models will help a lot. Youtube should have replicated servers in major market. As more players get in the video game I'm sure they will be setting up shop in several areas. Video doesn't change that much so when one person uploads it can be replicated throught out the network. You can still host the main links from a centralized place, but then stream the video from the closest location as it becomes available. This takes all the traffic from the west coast and keeps it there keeping people from the midwest from saturating the big pipes that connect the regions. Less hops also means less latency which is good for everyone.
People have been saying this same thing for ever. Telecom companies are just afraid of admitting that they can't charge up the ying yang for DS3s anymore. They are also going to have to invest in their networks which there shareholders hate. It is also the local telcoms that irritates me. Although dealing with Sprint is no treat, dealing with SBC/ATT/other momma bells is huge pain.
Networks are distributed by nature, so it just means you can't pipe all the data thru centralized routers. You are going to have to setup an infrastrute that can do very basic routing in a spider web. You can route packets very quickly if you just look at the first octect...and forward along to another router. All 1.xxx.xxx.xxx thru 5.xxx.xxx.xxx.xxx can be piped to a router that knows about those routes, and even breaks it down further. If you think about it they don't even need to do that they can just take the packet and load balance to many other devices. I think it'll be a while before we can't route faster...it is not like faster switching rates is completely dead.
If anything video is just forcing the issue of increasing the capacity, which will always need to grow. Eventually we will be streaming high end video content, and this article will be a long forgotten joke.
Yes Google is offering you to not use gmail, which is now a mute point, but not at the time.
4) Internal email does not require the internet. If I send an email to a coworker he will still get it even if our internet connection is down. In larger organization email is tranmitted over private lines. This does have some security merrit.
7) I'm sure google could offer faxes in the future with charge backs. Setting up an enterprise fax solution and managing numbers is a real pain. Google has the resources to setup it up and even give out numbers. Hopefully they are reading this:)
8) A public folder is not just a shared account. Each folder has permissions from read/write to creating directories. A shared account might be able to pull it off to some extent, but for us it was absolutely needed to manage the large number of blanket emails. And the shared password isn't PCI compliant.
Back when we where considering going from Exchange 5.5 to 2003 ( a huge pain in the butt ) I considered moving us to an online alternative. intranets.com now WebOffice ( webex umbrella ) provided somewhat of an alternative at that point. Now they are even better that they offer email hosting, with your domain not "gmail.com".
Several factors stopped me from being able to make that jump.
1) Legacy...everyone was using Exchange and we had tons of email in it that would be a pain to copy into folders. 2) Regulation. How does google keep all company emails in one place that can be archived and backed up. I'm sure Google won't loose someone's email anytime soon ( less likely then us ), but how do you document their backup procedures. 3) Current email addresses. No one wanted to give them up. 4) Internet bandwidth and reliance. People tend to think of the internet like electricity, but we are not there yet. It is funny that I get a faster connection at my house with a cable modem then our dual t1s provide...and a lot cheaper. This is another post, but unless you are in a big data center getting a decent sized pipe at a reasonable price is still overpriced. 5) Gateway level controls. We wanted to see every email that came in. We run a spam firewall, but if it blocks errantly we have a log. If Google blocks and email? 6) Customer support emails. We have tons of email addresses for our clients/etc that would probably be a pain to setup. 7) Fax support. We have to integrate with a fax server...yep it sucks. 8) Public folders ( ie email boxes accessible by more then one person )...ties in with 6.
To name a few.
If I was starting up a small software company I'd be all over this. As far as for enterprise uses...I think Google has a long road ahead of them...but they are speeding car.
It sounds all fine and dandy to allow the user to install all kinds of stuff on there machines. And without a company mandate with some teeth ( termination or write ups ) most people will install things on their own anyways. We have prevented people from having root access, but generally they figure out what the password is or someone in IT tells them.
The only problem with these sorts of users is the support they require when it turns out they don't know what they are doing. Any boob can install iTunes, but even the smarter ones start having problems trying to figure out why there machine crashes afterwords. Then IT is called and blamed.
I'm fine with having these users install whatever they want, just as long as they realize that when they have a problem of any kind of size ( word won't start ) I'm going to blast the machine. If they are smart enough to install all the extra software they are smart enough to put their data on the network or at least in one folder where I can copy it. If they say I lost all my MP3's I'm not going to have a problem telling them tough.
These same people don't have to sign the invoices for their expensive laptops, I do. It is company property and companies should have every right to tell individuals what they can and can't install. At the same time they cannot be so stubborn as to not allow for newer software to get added, even if it does pose some sort of risk. Instant messenger and those types of programs can greatly increase productivity if used correctly. If the employee is chatting with his wife, I'd rather he do that then go in the hallway and call him on his cell...chances are he is actually doing something in between the chat lines.
That said the company still has the right to monitor the person for any traffic going over their network. If the guy gets in trouble and they find that he chatted with his wife all the time it should be admissable in determining his dismisal. Everyone out there knows when enough is enough, those that don't usually end up without a job.
Say HD DVD to someone and get them to guess what it means. A large percentage will at least know that it is some kind of DVD.
Now say Blu-ray to another set of people. I doubt anyone will have a clue. They will likely think it is some kind of wireless network because of the blutooth branding.
I think it will have a much faster adoption because the physical medium and playback are so similiar. As soon as the players get to around 150-200 and they play all three types then people will buy the HD-DVDs at best buy over the blue ray because it makes sense and sony will give up. There will be a version of the PS3 with a hybrid drive in less then 3 years I reckon.
The exact amount of time? Have you ever played a musical instrument. It sounds like you have, but I don't know why you say this.
I've actually tried to play the guitar before playing guitar hero. I also played trumpet for about 10 years.
Guitar hero is not supposed to be a guitar simulation. It is more of a rock and roll simulation. And if you played it on expert it can be quite challenging and rewarding. It is a game like any other...so stop thinking so much and STFU.
The porn industry is not a stupid and corrupt as it might appear. They don't want little kids staring at this anymore then parents do. They are not the tobacco industry. They want to protect their interests which means not showing it to kids which is against the law.
They will switch over because it make them look legit. Someone sees the XXX domain and knows that it will probably be a good porn site. I wish there where actually some laws passed to protect the consumer.
Say I get some malware after visiting one of the sites. A repersentative from.XXX will review and then take away the domain name if they do not stop.
I'd say make them pay $1000 dollars for the domain name. With yearly fees. People will pay it if there is money, and the consumer will pay for fast downloads and good security.
The companies that don't switch over will still exist for sure. So blocking them will not be possilbe without the use of a service, but I really don't care. I want a designated area on the net for porn...and this will give us it.
Windows really should have put out a new build of XP before releasing Vista. Just SP2 with a new installer that mimics Windows server 2003. If you've ever installed Windows Server 2003 it can be quite secure. It turns off all inbound connections until you can install patches. It turns off IE so you can't surf anything without explicity telling it you are ready to. Server 2003 was going down the right path, I'm not sure why they never ported some of these basics to a new XP back in 2004. I guess it's too late now.
The last big Windows worm was quite a while ago. They are still alive thanks to the unaware. Windows has a lot of ports open compared to other machines mostly because it was designed to operate in a operate in an Active directory enviornment...and because RPC is overally relied upon. Yes you can get a virus delivered by email, but this is true of any OS where the user is running as root ( admin ( if the os even supports it ) ) and opens up an attachment. Windows users are bombared with viruses that Mac users get and can safely ignore...heck if you tried to run the exe it would just fail. Mail virsuses are getting less and less as well as email providers and spam firewalls are blocking them. A properly written virus ran on Linux or Mac OSx can get thru the protection. Linux and Mac OSx have had plenty of exploits to get a file install things.
While other OSes interact with each other, they don't quite do it with the built in way MS does. This is good for the end user and bad for security. SMB setup has gotten a heck of lot easier on Linux in the last few years, but compared to Windows it'll never be quite as easy. There are products out there like Groupware, but Active Directory is by far the simplest and most useful for setting up a small to massive network. Thousands of companies use it every day to share files and get work done. Install a printer from the active directory isn't super easy, but I ca'tn see a Linux product comparing.
Mac interaction with AD isn't that bad. I wish it had an Active Directory client from the get go, but my Mac users can print, share files, and a few other things okay. Nobody likes to mention that Windows file security is far more advanced then Linux's will be for quite sometime. The ability to permission a file to individual users at varying levels is absolutely crucial. It is a pain for my Mac users to have to remember their NT passwords and visit a NT machine to reset it every once in a while, but it is good enough so they can run Photoshop...with the Mac keyboard.
I won't be suprised to see a mac mode in Vista sometime soon. It wouldn't really be that hard for Windows to stick the file menu up on the top of the screen when a Window takes focus.
The fact of the matter that no ones wants to talk about is MS is becoming fairly secure if installed with it's patches and stuck behind a firewall. This is true of practically all OSes. The big problem MS has it that it doesn't update it's install disks and most of it's vendors don't update their freaking images. If I get a new Dell I would expect not to have to install a single patch that was over two months old, but alas they don't do that for you. Imagie you installed Redhat 3.0 and then put yourself on the network. I'm sure someone out there could right a worm for Redhat 3.0 right? There isn't one port in the default install with a buffer overflow issue? It be an interesting expierement to write worms for older versions of OSs and see how they take. My guess is that there are more Windows 98 boxes running today then RedHat 3.0 boxes ever ran.
The point is OSx or Linux get the marketshare that Windows has you'll see 1000's of older versions of the OS. As it sicks MAC users generally upgrade fast, and Linux users are practically religous about it outside of the server scope. And on the server side it is likely the machines are protected via firewalls.
The browser hole is getting plugged as we speak. Firefox, Opera, and IE are all plugging away. The big issues is that HTML and Javascript t
As has been stated many times over a plug-in car just offset the polution to our power generators...which are already overloaded.
I think what we really need to ask ourselves is if we need a 300 hp beast when we go to the grocery store. Pistons that automatically shut down when not need. A solar panel that might get you a few miles on the way home by charging a battery during the day.
I was down in Dallas for work back in the summer. Everyone drove, but everyone also parked outdoors. The sun light was intense and it was hot as heck. All the grass was burnt. Solar panels on the top of the cars and on the burnt lawns ( dust circles ) would have been enough to save quite a bit of gas over the weekend.
If you live in a windy area could we not have little mini generators on your car when it is parked? Even if a full day of wind only provided 2 minutes of driving electricty is it not worth it?
If we are going electric until we can figure out the hydrogen thing ( if ever ) we need to be more creative on how we charge them. Just plugging in the car won't solve the problem.
One of the biggest issues I have with Ajax and really what the web browser has turned into is Javascript.
Don't get me wrong, I think the language is "alright". The problem is it is the only choice out there.
There are many flavors of Linux and pratically everytype of appliation out there. But only one real choice for scripting on a web page.
Ajax's primary function is the ability to grab content from a web server ( or any server ) and then modify DOM of the web page you are working on.
The could be done by any programming language. Most of the time I use Ajax I just grab some HTML content and blast it into a div or some other object. Here Jscript isn't that big of a deal with a good library.
When I start to have to actually grab the value and parse it that is where JScript is the only option. I wish that browser would stat offering some choice. Maybe Ruby or Python or Perl, don't really care. Just something else besides JScript. Seems like ActionScript might get a chance.
I find it funny what is happening here. I think we are forgetting that the internet is just a medium to exchange data, not an application.
The web browser is not the internet. The web browser is an application which read html and javascript instead of byte code to execute. Google's new spreadsheet thing is just a big javascript app that happens to save it's data on a server.
MS is behind in that SharePoint is a piece of crap. They have not figured out how to replace the save dialog with one that works well. MS need Excel Server 2007. I don't have care if I have to install Office 2007 in order to access it, it has nothing to do with the client. In fact I'd prefer a thick client with some sort of caching so I could work on it from home.
I've been Web developing for a while, and quite frankly the best internet enabled apps are still C++ apps that talk to the internet via HTTP or some other protocol. I wouldn't mind a google that was full bonified install this app.
In the long run Ajax can only go so far. C++ is just a much more robust language then javascript will ever be. What I think we will see if Firefox as a module that you can manipulate with C++ code. DHTML is proving to be a great layout language...minus forms. I really don't think javascript is scaling...it works, but it only seems hacky to me.
I'm not to worried about this because most users are aware of attachment exploits like this.
I'm sure the major spam firewalls will also have signatures in a relatively short period of time. If my email spam/virus firewall will stop this I'm fine.
For the home user it is a bit more of an issue. At the same time most people use Yahoo, MSN, Google or some other account that has active scanner that I'm sure will be able to block these in the short run...if not by analyzing the file by analyzing the subject line. Heck, chances are it'll look like spam to my firewall won't let it thru to begin with.
I do wish MS would put out the technical details of this exploit. It sounds like some sort of a buffer overflow. Something tells me it is a graphic insert of some sort, but who knows.
Why don't we get the day off. I was at work late last night, and I'm behind on some projects. I barely have time to take a lunch let alone go stand in a line for an hour. We get all kinds of silly days off in the United States. President's day!!! But not election day. I'm very disappointed in our reps on both side of the isle in dealing with the voting situation.
As for paper ballots I think we should stick with them until we get a system ironed out. At the same time they are not perfect either. Remember the Buccanon debacle in the Florida 00 election.
I'd like to see each voter get a random "card" with a bar code on it. This would be unique for everyone and handed out randomly at the polling station. Then you would stick that card into a machine which would record your vote and the bar code. Then later you could go online and scan it in...or some office...and "verify" your vote. Furthermore I think we should use two different system from two different vendors. Even better to have the Republicans choose one and the Democrats the other. Then when the country goes to verify the vote they can make sure that both machines match up.
When you walk away from the machine(s) you should get a paper copy that you can use to double verify. If we can spend 100's of billions on war, I think we can spend some cash on our election systems.
"Cryptic Registry Settings - I've never quite gotten why it was determined that putting all your settings and configuration in one basket was deemed to be a good idea. I can't think of any positive justification whatsoever for this."
The concept to me is far better then an etc folder. Why, because there is one consistent format. Each file in an etc can be completely different. I can search the entire registry with one easy tool, versus having to use a tool to search through each file.
Even on Windows I have programs that don't use the registry...at the end of the day it is a techy or developer tool. The end user really doesn't care, and both seem to take care of the problem.
Now the reason everyone hates the registry and the reason it sucks on Windows is because it doesn't get defragmented...and gets fragged pretty badly...and MS did a crappy job of indexing, programs get full access, and so on and so on. It is the implementation more then the concept itself IMHO.
Slashdot posted to a story the other day that has a lot to do with MS woes. That was the story about MS not forcing their own employees to run as non-admins.
The home version actually seems to do this a little better then the Professional edition. My parents and my siblings login to their computer with their own accounts and all as non-admins. Of course they don't lock down admin, so they all install all kinds of stuff.
Vista seems to go a long way in stopping unwanted stuff from installing, but with such a mainstream system does is it really going to help? If a user has to switch to admin to install that screensaver that also is spyware does that really help? Does Ms have to be held accountable for Spyware that is purposely installed on a machine. If it comes in through IE 6...sure it is MSs fault.
Linux installers are applauded by most, but I wonder what will happen to them in the mainstream. Commercial software will probably still install with stand-alone installers if Linux where to take off. Linux ( and others ) has standards that adhered to via open source packages, but would another company really put up with it. So a user in Linux goes to run an executable off a web page...they get an error from it saying please be in root mode. If they login as root would Linux do anything to stop them from overwriting system config files? Would we blame the problem on Linux or the author?
The author seems to be misplacing the blame. MS has to be the app cops? I guess in this day and age yes...5 years ago...not so much.
In the long run I think all OS's need to force application to install in virtual file systems. When I go to install a major app I wish that it would just copy a big file and "mount" it to the machine. You wouldn't even need to be in root to do it if done via an API call. The app would be registered with the OS and given a small amount of hardrive space to write it's config files to that only it would have access to. When it goes to save data files for the user the OS would ask the user if it was alright for it to. We can run entire OSs in a VMWare like system, why not applications themselves.
Of course lots of apps, especially in OS use pipes and heavily rely on other systems and libraries. Back in the day when sharing a DLL was needed to save HD space it was a good idea...is it now. Should we require all the apps to include their libraries? This would make code injects a lot harder as well....sorry botters.
The fundamental idea of an App installing needs to be re-engineered. Some OSs do a better job then others, but they all fundamentally invovled the installer coping files around, which will always lead to the types of problems we are seeing.
Has a Red Hat, Suse, Debian build ever come out bug free. Hard to say since so many of the packages that you can install via apt or whatever are not really associated. If apache has a bug it's apache's fault, not Windows.
This is a major disadvantage, but also a major advantage that both Windows and Mac to some extent share.
With any software you have to get it out the door. It'll never be perfect, and no matter how long they wait there will be an SP1 fairly soon.
To me what RC1 means is that nothing big and fancy is going to get ADDED. What you see if pretty much what you get. If a major flaw is found they might rearrange a piece of functionality, but most things are going to be bug fixes.
While in Beta they might completely take something out. In RC you probably are not going to get away with it, although you migth "delay" something to SP1 like Microsoft did with database mirroring in SQL 2005 in order to get it out the door.
As much as I hate patching, I'd rather get it out in the field and get some use out of it. Early adopters will get hit the hardest, but that is what they expect. Dell and the other manufactors will be the ones finding most of the bugs from now on anyways.
It isn't easy, but one day I think we will have a good database based file system. I don't think we need to throw away our good old directory based system for our system files and other static files, but I do think it'd be good for data and perhaps configs.
If you've used a major database you'll know they use a singular or multiple files which contain all of the data. The data is usually seperated into 2/4/8 kb pages depending on the system. When you do a backup the pages are copied into a backup file. The backup size is 8kb times the number of pages used roughly. When each page is backed up it is tagged as being so, this makes differentials a breeze. When a page is written to, the backup stamp is removed and then the backup software can just backup those pages to make a differential backup. Differentials grow with time, so you'll do full backups every week or so and then diffs each week.
Tag on a transaction log file and you can back it up at the end of the main backup and have real time backups.
I'd love to just save my mp3s, videos, and other static data files into a "repository" and just search for them as I needed them. One of the big challenges to me is re-writing all the apps out there to take advantage of such a system, especially with the classic file model being so engrained in the basics of most operating systems.
Windows has promised this for many years, but I'm guessing it will be a brand new kid on the block that will do that. Maybe if Google write an OS from scrath or something of that nature...pretty far fetch though.
As I was reading over this I got that idea that maybe we could do the same thing, but with our clusters on the harddrive. If HD manufacturs would add a little bit on the harddrive itself we might be a to do a block level differential.
When things like this come out; things like key checking for a game install and everything else that is designed to stop piracy I often wonder who wrote it?
Are the best and brightest out there the ones that get stuck with this task? I would think it'd be the interns and that developers everyone hates that get the fun task.
I've used products that had good licensing tools. Keys that you enabled online, and enabled a number of users etc. Everytime it seems like it comes out of some smaller software company with small bright teams. I'm guessing in these cases the senior level codes and maybe even the whole team got involved.
Anyone out there have expierence writing key checkers and other piracy related pieces of functionality?
Wow what a rant, that got up to 5 ( Insightful ). Way to go slashdot.
HTML/CSS/JavaScript like any technology is getting old. It wasn't designed to really be for applications. Now we have Ajax hacks and a slew of other crap to try and make it like a normal desktop app...things that flash and java applets ( yes I know applets are not that great ) just do.
Flash can be just as accesible if not more then a web page...it is all in the tools that make it accesible. Imagine if I wrote a flash app specifically for blind people...I'm guessing I could get a lot further then with just a web page. Instead of trying to make a page accessible...i'd rather see a version of the app written specifically for blind people. It'd be better if google or other companies teamed up with another company, give them the raw content as XML and let them expose it in a way that will make it easier to access. Browser are inherently visual are they not....maybe it'd make more sense for google to try and expose the information in a way that could be converted to brail or audio easily.
I'm sorry to the individuals out there that have disabilities. At the same time some content that is very hard if not impossible to make accessible can make it far easy to access for people without a disablity to use it. We need to find ways to appease both communities.
What does the poster mean by the "real Web". And they SHOULD just post PDFs on the side of their content, they are way more accesible then HTML. I mean I'm sure you could get a program to read the content of a PDF far easier then you could get it to strip out text from an html document and read it.
As much as I think this could help in the event of disaster, I hope that some legislation is passed to limit it's use. It would be very easy to abuse it for propogand purposes.
Funny that Web 2.0 is taking off so much. The problem with it is that everyone I interview is now "learning" Ajax. I feel like if I go to an interview I'll be asked a million Ajax questions, that I really don't want to answer.
Using hidden Iframes and JScript was one way to do what Ajax does years ago. There are definately a few cases where it is really useful. A little div popup, pre-populating city state after a postal code was entered, testing a value etc. Debugging is much harder, and the Javascript/DOM model is hard to code bug free. Javascript errors don't get reported to the server admin, and they are often hard to replicate. This is partly a lack of good tools, but view source on HTML is almost always easier then trying to step thru some buggy jscript.
It can be very easy to abuse Ajax. I recently had someone show me a search example that "pre-populated" as you typed. It was super clunky and really didn't work. Ajax's biggest problem at this point is that everyone thinksd everything has to be instant now. You can make a user go to another page to edit something that is not edited every other minute.
As much as I love Google maps, Yahoo Flash maps kick their ass. Adobe's new Flex tech is really going to give Ajax a run for the money. Java is just to sluggish, but Flash is pretty quick. Yes you'll have to turn off your flash ad blockers.
The thing that has to happen is that SVG or a new standard needs to be born to handle GUI apps. People don't like flash because there is a name behind it, HTML is a standard, Javascript is a standard, etc. Java is Sun/IBM, Flash is Adobe ( formally Macromedia ).
I guess my biggest gripe with Web 2.0 is that almost everything that we spend hours figuring out in JScript could be done if people would create more and better HTML tags. Then the browser developers take care of all the testing, and we will have more stable apps.
Personally I'm going down the Flash path. If you haven't tried Flex yet, labs.adobe.com, do yourself a favor and see what you've been missing....no I don't work for Adboe or even really like them:) You can do more in less time, and you can create content that really looks good. I'd love to see a Flex slashdot version.
All of these ideas are old, and while high performing don't address the largest issue of all, cross kernel compatability.
Sure you can recompile and all that jaz, but I'd love to see a day where an app could run on any number of kernels out there. This creates real competetion.
What I'd like to see if a kernel more like a CPU. Instead of linking your kernel calls, you place them as if you where placing an Assembly call. Then we can have many companies and open source organizations writing versions of it.
As we move towards multi core cpus this could really lead to performance leads. Where one or more of many cores could be dedicated to the kernel operations listening for operations and taking care of them. No context switches needed, no privledge mode switching.
Drivers and everything else run outside of kernel mode and use low level microcode to execute the code.
The best part I think is you could make it backword compatiable as we re-write. A layer could handle old kernel calls and change them to the micro codes.
As we define everything more and more then we might even be able to design CPUs that can handle it better.
Slashdot Mirror
This is a flaw of computers....yes it results from bad coding...but should a simple mistake allow for code to be executed.
Why doesn't the no execute bit fix this?
The internet will continue to grow in capacity and as it has new products will come out to fill the void.
My biggest issues with youtube are at work in our main office. We have a large application hosted in data center. It is a major hub for internet connectivity for the region. Given that we are so close to some big vendors we can get lots of bandwith for relatively low prices. If my employees where sitting in that facility they could surf youtube.com all day.
Now at home I can also do it. I pay Comcast a big more for the extra bandwidth and I can download over a meg a second from some sites. Verizon is going to be laying fiber directly to houses and businesses soon.
Get into our offices and it is a different story. We have dual t1s coming in and only 60+ employees, but we are constantly saturated. Combine that with the fact that Cisco Pixes have horrible throttling support and you end up with times when I can't even access basic websites very quickly. The issue here is that T1s and DS3s are freakin expensive compared to a simple cable modem. We have been tempted to get Comcast bussines ( which makes me shiver a bit ) because I can get larger down pipes for general internet surfing. We only host a few services such as email here so it isn't like we need megs of up bandwith.
Throttling would go along way to solve this issue. Youtube could buffer people down quite a bit, you would just have to wait for the movie to buffer a bit. For shared internet connections and ISPs this could allow for better QOS.
Distribution models will help a lot. Youtube should have replicated servers in major market. As more players get in the video game I'm sure they will be setting up shop in several areas. Video doesn't change that much so when one person uploads it can be replicated throught out the network. You can still host the main links from a centralized place, but then stream the video from the closest location as it becomes available. This takes all the traffic from the west coast and keeps it there keeping people from the midwest from saturating the big pipes that connect the regions. Less hops also means less latency which is good for everyone.
People have been saying this same thing for ever. Telecom companies are just afraid of admitting that they can't charge up the ying yang for DS3s anymore. They are also going to have to invest in their networks which there shareholders hate. It is also the local telcoms that irritates me. Although dealing with Sprint is no treat, dealing with SBC/ATT/other momma bells is huge pain.
Networks are distributed by nature, so it just means you can't pipe all the data thru centralized routers. You are going to have to setup an infrastrute that can do very basic routing in a spider web. You can route packets very quickly if you just look at the first octect...and forward along to another router. All 1.xxx.xxx.xxx thru 5.xxx.xxx.xxx.xxx can be piped to a router that knows about those routes, and even breaks it down further. If you think about it they don't even need to do that they can just take the packet and load balance to many other devices. I think it'll be a while before we can't route faster...it is not like faster switching rates is completely dead.
If anything video is just forcing the issue of increasing the capacity, which will always need to grow. Eventually we will be streaming high end video content, and this article will be a long forgotten joke.
Yes Google is offering you to not use gmail, which is now a mute point, but not at the time.
:)
4) Internal email does not require the internet. If I send an email to a coworker he will still get it even if our internet connection is down. In larger organization email is tranmitted over private lines. This does have some security merrit.
7) I'm sure google could offer faxes in the future with charge backs. Setting up an enterprise fax solution and managing numbers is a real pain. Google has the resources to setup it up and even give out numbers. Hopefully they are reading this
8) A public folder is not just a shared account. Each folder has permissions from read/write to creating directories. A shared account might be able to pull it off to some extent, but for us it was absolutely needed to manage the large number of blanket emails. And the shared password isn't PCI compliant.
Back when we where considering going from Exchange 5.5 to 2003 ( a huge pain in the butt ) I considered moving us to an online alternative. intranets.com now WebOffice ( webex umbrella ) provided somewhat of an alternative at that point. Now they are even better that they offer email hosting, with your domain not "gmail.com".
Several factors stopped me from being able to make that jump.
1) Legacy...everyone was using Exchange and we had tons of email in it that would be a pain to copy into folders.
2) Regulation. How does google keep all company emails in one place that can be archived and backed up. I'm sure Google won't loose someone's email anytime soon ( less likely then us ), but how do you document their backup procedures.
3) Current email addresses. No one wanted to give them up.
4) Internet bandwidth and reliance. People tend to think of the internet like electricity, but we are not there yet. It is funny that I get a faster connection at my house with a cable modem then our dual t1s provide...and a lot cheaper. This is another post, but unless you are in a big data center getting a decent sized pipe at a reasonable price is still overpriced.
5) Gateway level controls. We wanted to see every email that came in. We run a spam firewall, but if it blocks errantly we have a log. If Google blocks and email?
6) Customer support emails. We have tons of email addresses for our clients/etc that would probably be a pain to setup.
7) Fax support. We have to integrate with a fax server...yep it sucks.
8) Public folders ( ie email boxes accessible by more then one person )...ties in with 6.
To name a few.
If I was starting up a small software company I'd be all over this. As far as for enterprise uses...I think Google has a long road ahead of them...but they are speeding car.
It sounds all fine and dandy to allow the user to install all kinds of stuff on there machines. And without a company mandate with some teeth ( termination or write ups ) most people will install things on their own anyways. We have prevented people from having root access, but generally they figure out what the password is or someone in IT tells them.
The only problem with these sorts of users is the support they require when it turns out they don't know what they are doing. Any boob can install iTunes, but even the smarter ones start having problems trying to figure out why there machine crashes afterwords. Then IT is called and blamed.
I'm fine with having these users install whatever they want, just as long as they realize that when they have a problem of any kind of size ( word won't start ) I'm going to blast the machine. If they are smart enough to install all the extra software they are smart enough to put their data on the network or at least in one folder where I can copy it. If they say I lost all my MP3's I'm not going to have a problem telling them tough.
These same people don't have to sign the invoices for their expensive laptops, I do. It is company property and companies should have every right to tell individuals what they can and can't install. At the same time they cannot be so stubborn as to not allow for newer software to get added, even if it does pose some sort of risk. Instant messenger and those types of programs can greatly increase productivity if used correctly. If the employee is chatting with his wife, I'd rather he do that then go in the hallway and call him on his cell...chances are he is actually doing something in between the chat lines.
That said the company still has the right to monitor the person for any traffic going over their network. If the guy gets in trouble and they find that he chatted with his wife all the time it should be admissable in determining his dismisal. Everyone out there knows when enough is enough, those that don't usually end up without a job.
Say HD DVD to someone and get them to guess what it means. A large percentage will at least know that it is some kind of DVD.
Now say Blu-ray to another set of people. I doubt anyone will have a clue. They will likely think it is some kind of wireless network because of the blutooth branding.
I think it will have a much faster adoption because the physical medium and playback are so similiar. As soon as the players get to around 150-200 and they play all three types then people will buy the HD-DVDs at best buy over the blue ray because it makes sense and sony will give up. There will be a version of the PS3 with a hybrid drive in less then 3 years I reckon.
The exact amount of time? Have you ever played a musical instrument. It sounds like you have, but I don't know why you say this.
I've actually tried to play the guitar before playing guitar hero. I also played trumpet for about 10 years.
Guitar hero is not supposed to be a guitar simulation. It is more of a rock and roll simulation. And if you played it on expert it can be quite challenging and rewarding. It is a game like any other...so stop thinking so much and STFU.
The porn industry is not a stupid and corrupt as it might appear. They don't want little kids staring at this anymore then parents do. They are not the tobacco industry. They want to protect their interests which means not showing it to kids which is against the law.
.XXX will review and then take away the domain name if they do not stop.
They will switch over because it make them look legit. Someone sees the XXX domain and knows that it will probably be a good porn site. I wish there where actually some laws passed to protect the consumer.
Say I get some malware after visiting one of the sites. A repersentative from
I'd say make them pay $1000 dollars for the domain name. With yearly fees. People will pay it if there is money, and the consumer will pay for fast downloads and good security.
The companies that don't switch over will still exist for sure. So blocking them will not be possilbe without the use of a service, but I really don't care. I want a designated area on the net for porn...and this will give us it.
Windows really should have put out a new build of XP before releasing Vista. Just SP2 with a new installer that mimics Windows server 2003. If you've ever installed Windows Server 2003 it can be quite secure. It turns off all inbound connections until you can install patches. It turns off IE so you can't surf anything without explicity telling it you are ready to. Server 2003 was going down the right path, I'm not sure why they never ported some of these basics to a new XP back in 2004. I guess it's too late now.
The last big Windows worm was quite a while ago. They are still alive thanks to the unaware. Windows has a lot of ports open compared to other machines mostly because it was designed to operate in a operate in an Active directory enviornment...and because RPC is overally relied upon. Yes you can get a virus delivered by email, but this is true of any OS where the user is running as root ( admin ( if the os even supports it ) ) and opens up an attachment. Windows users are bombared with viruses that Mac users get and can safely ignore...heck if you tried to run the exe it would just fail. Mail virsuses are getting less and less as well as email providers and spam firewalls are blocking them. A properly written virus ran on Linux or Mac OSx can get thru the protection. Linux and Mac OSx have had plenty of exploits to get a file install things.
While other OSes interact with each other, they don't quite do it with the built in way MS does. This is good for the end user and bad for security. SMB setup has gotten a heck of lot easier on Linux in the last few years, but compared to Windows it'll never be quite as easy. There are products out there like Groupware, but Active Directory is by far the simplest and most useful for setting up a small to massive network. Thousands of companies use it every day to share files and get work done. Install a printer from the active directory isn't super easy, but I ca'tn see a Linux product comparing.
Mac interaction with AD isn't that bad. I wish it had an Active Directory client from the get go, but my Mac users can print, share files, and a few other things okay. Nobody likes to mention that Windows file security is far more advanced then Linux's will be for quite sometime. The ability to permission a file to individual users at varying levels is absolutely crucial. It is a pain for my Mac users to have to remember their NT passwords and visit a NT machine to reset it every once in a while, but it is good enough so they can run Photoshop...with the Mac keyboard.
I won't be suprised to see a mac mode in Vista sometime soon. It wouldn't really be that hard for Windows to stick the file menu up on the top of the screen when a Window takes focus.
The fact of the matter that no ones wants to talk about is MS is becoming fairly secure if installed with it's patches and stuck behind a firewall. This is true of practically all OSes. The big problem MS has it that it doesn't update it's install disks and most of it's vendors don't update their freaking images. If I get a new Dell I would expect not to have to install a single patch that was over two months old, but alas they don't do that for you. Imagie you installed Redhat 3.0 and then put yourself on the network. I'm sure someone out there could right a worm for Redhat 3.0 right? There isn't one port in the default install with a buffer overflow issue? It be an interesting expierement to write worms for older versions of OSs and see how they take. My guess is that there are more Windows 98 boxes running today then RedHat 3.0 boxes ever ran.
The point is OSx or Linux get the marketshare that Windows has you'll see 1000's of older versions of the OS. As it sicks MAC users generally upgrade fast, and Linux users are practically religous about it outside of the server scope. And on the server side it is likely the machines are protected via firewalls.
The browser hole is getting plugged as we speak. Firefox, Opera, and IE are all plugging away. The big issues is that HTML and Javascript t
As has been stated many times over a plug-in car just offset the polution to our power generators...which are already overloaded.
I think what we really need to ask ourselves is if we need a 300 hp beast when we go to the grocery store. Pistons that automatically shut down when not need. A solar panel that might get you a few miles on the way home by charging a battery during the day.
I was down in Dallas for work back in the summer. Everyone drove, but everyone also parked outdoors. The sun light was intense and it was hot as heck. All the grass was burnt. Solar panels on the top of the cars and on the burnt lawns ( dust circles ) would have been enough to save quite a bit of gas over the weekend.
If you live in a windy area could we not have little mini generators on your car when it is parked? Even if a full day of wind only provided 2 minutes of driving electricty is it not worth it?
If we are going electric until we can figure out the hydrogen thing ( if ever ) we need to be more creative on how we charge them. Just plugging in the car won't solve the problem.
One of the biggest issues I have with Ajax and really what the web browser has turned into is Javascript.
Don't get me wrong, I think the language is "alright". The problem is it is the only choice out there.
There are many flavors of Linux and pratically everytype of appliation out there. But only one real choice for scripting on a web page.
Ajax's primary function is the ability to grab content from a web server ( or any server ) and then modify DOM of the web page you are working on.
The could be done by any programming language. Most of the time I use Ajax I just grab some HTML content and blast it into a div or some other object. Here Jscript isn't that big of a deal with a good library.
When I start to have to actually grab the value and parse it that is where JScript is the only option. I wish that browser would stat offering some choice. Maybe Ruby or Python or Perl, don't really care. Just something else besides JScript. Seems like ActionScript might get a chance.
I find it funny what is happening here. I think we are forgetting that the internet is just a medium to exchange data, not an application.
The web browser is not the internet. The web browser is an application which read html and javascript instead of byte code to execute. Google's new spreadsheet thing is just a big javascript app that happens to save it's data on a server.
MS is behind in that SharePoint is a piece of crap. They have not figured out how to replace the save dialog with one that works well. MS need Excel Server 2007. I don't have care if I have to install Office 2007 in order to access it, it has nothing to do with the client. In fact I'd prefer a thick client with some sort of caching so I could work on it from home.
I've been Web developing for a while, and quite frankly the best internet enabled apps are still C++ apps that talk to the internet via HTTP or some other protocol. I wouldn't mind a google that was full bonified install this app.
In the long run Ajax can only go so far. C++ is just a much more robust language then javascript will ever be. What I think we will see if Firefox as a module that you can manipulate with C++ code. DHTML is proving to be a great layout language...minus forms. I really don't think javascript is scaling...it works, but it only seems hacky to me.
I'm not to worried about this because most users are aware of attachment exploits like this.
I'm sure the major spam firewalls will also have signatures in a relatively short period of time. If my email spam/virus firewall will stop this I'm fine.
For the home user it is a bit more of an issue. At the same time most people use Yahoo, MSN, Google or some other account that has active scanner that I'm sure will be able to block these in the short run...if not by analyzing the file by analyzing the subject line. Heck, chances are it'll look like spam to my firewall won't let it thru to begin with.
I do wish MS would put out the technical details of this exploit. It sounds like some sort of a buffer overflow. Something tells me it is a graphic insert of some sort, but who knows.
Why don't we get the day off. I was at work late last night, and I'm behind on some projects. I barely have time to take a lunch let alone go stand in a line for an hour. We get all kinds of silly days off in the United States. President's day!!! But not election day. I'm very disappointed in our reps on both side of the isle in dealing with the voting situation.
As for paper ballots I think we should stick with them until we get a system ironed out. At the same time they are not perfect either. Remember the Buccanon debacle in the Florida 00 election.
I'd like to see each voter get a random "card" with a bar code on it. This would be unique for everyone and handed out randomly at the polling station. Then you would stick that card into a machine which would record your vote and the bar code. Then later you could go online and scan it in...or some office...and "verify" your vote. Furthermore I think we should use two different system from two different vendors. Even better to have the Republicans choose one and the Democrats the other. Then when the country goes to verify the vote they can make sure that both machines match up.
When you walk away from the machine(s) you should get a paper copy that you can use to double verify. If we can spend 100's of billions on war, I think we can spend some cash on our election systems.
"Cryptic Registry Settings - I've never quite gotten why it was determined that putting all your settings and configuration in one basket was deemed to be a good idea. I can't think of any positive justification whatsoever for this."
The concept to me is far better then an etc folder. Why, because there is one consistent format. Each file in an etc can be completely different. I can search the entire registry with one easy tool, versus having to use a tool to search through each file.
Even on Windows I have programs that don't use the registry...at the end of the day it is a techy or developer tool. The end user really doesn't care, and both seem to take care of the problem.
Now the reason everyone hates the registry and the reason it sucks on Windows is because it doesn't get defragmented...and gets fragged pretty badly...and MS did a crappy job of indexing, programs get full access, and so on and so on. It is the implementation more then the concept itself IMHO.
Slashdot posted to a story the other day that has a lot to do with MS woes. That was the story about MS not forcing their own employees to run as non-admins.
The home version actually seems to do this a little better then the Professional edition. My parents and my siblings login to their computer with their own accounts and all as non-admins. Of course they don't lock down admin, so they all install all kinds of stuff.
Vista seems to go a long way in stopping unwanted stuff from installing, but with such a mainstream system does is it really going to help? If a user has to switch to admin to install that screensaver that also is spyware does that really help? Does Ms have to be held accountable for Spyware that is purposely installed on a machine. If it comes in through IE 6...sure it is MSs fault.
Linux installers are applauded by most, but I wonder what will happen to them in the mainstream. Commercial software will probably still install with stand-alone installers if Linux where to take off. Linux ( and others ) has standards that adhered to via open source packages, but would another company really put up with it. So a user in Linux goes to run an executable off a web page...they get an error from it saying please be in root mode. If they login as root would Linux do anything to stop them from overwriting system config files? Would we blame the problem on Linux or the author?
The author seems to be misplacing the blame. MS has to be the app cops? I guess in this day and age yes...5 years ago...not so much.
In the long run I think all OS's need to force application to install in virtual file systems. When I go to install a major app I wish that it would just copy a big file and "mount" it to the machine. You wouldn't even need to be in root to do it if done via an API call. The app would be registered with the OS and given a small amount of hardrive space to write it's config files to that only it would have access to. When it goes to save data files for the user the OS would ask the user if it was alright for it to. We can run entire OSs in a VMWare like system, why not applications themselves.
Of course lots of apps, especially in OS use pipes and heavily rely on other systems and libraries. Back in the day when sharing a DLL was needed to save HD space it was a good idea...is it now. Should we require all the apps to include their libraries? This would make code injects a lot harder as well....sorry botters.
The fundamental idea of an App installing needs to be re-engineered. Some OSs do a better job then others, but they all fundamentally invovled the installer coping files around, which will always lead to the types of problems we are seeing.
Has a Red Hat, Suse, Debian build ever come out bug free. Hard to say since so many of the packages that you can install via apt or whatever are not really associated. If apache has a bug it's apache's fault, not Windows.
This is a major disadvantage, but also a major advantage that both Windows and Mac to some extent share.
With any software you have to get it out the door. It'll never be perfect, and no matter how long they wait there will be an SP1 fairly soon.
To me what RC1 means is that nothing big and fancy is going to get ADDED. What you see if pretty much what you get. If a major flaw is found they might rearrange a piece of functionality, but most things are going to be bug fixes.
While in Beta they might completely take something out. In RC you probably are not going to get away with it, although you migth "delay" something to SP1 like Microsoft did with database mirroring in SQL 2005 in order to get it out the door.
As much as I hate patching, I'd rather get it out in the field and get some use out of it. Early adopters will get hit the hardest, but that is what they expect. Dell and the other manufactors will be the ones finding most of the bugs from now on anyways.
It isn't easy, but one day I think we will have a good database based file system. I don't think we need to throw away our good old directory based system for our system files and other static files, but I do think it'd be good for data and perhaps configs.
If you've used a major database you'll know they use a singular or multiple files which contain all of the data. The data is usually seperated into 2/4/8 kb pages depending on the system. When you do a backup the pages are copied into a backup file. The backup size is 8kb times the number of pages used roughly. When each page is backed up it is tagged as being so, this makes differentials a breeze. When a page is written to, the backup stamp is removed and then the backup software can just backup those pages to make a differential backup. Differentials grow with time, so you'll do full backups every week or so and then diffs each week.
Tag on a transaction log file and you can back it up at the end of the main backup and have real time backups.
I'd love to just save my mp3s, videos, and other static data files into a "repository" and just search for them as I needed them. One of the big challenges to me is re-writing all the apps out there to take advantage of such a system, especially with the classic file model being so engrained in the basics of most operating systems.
Windows has promised this for many years, but I'm guessing it will be a brand new kid on the block that will do that. Maybe if Google write an OS from scrath or something of that nature...pretty far fetch though.
As I was reading over this I got that idea that maybe we could do the same thing, but with our clusters on the harddrive. If HD manufacturs would add a little bit on the harddrive itself we might be a to do a block level differential.
When things like this come out; things like key checking for a game install and everything else that is designed to stop piracy I often wonder who wrote it?
Are the best and brightest out there the ones that get stuck with this task? I would think it'd be the interns and that developers everyone hates that get the fun task.
I've used products that had good licensing tools. Keys that you enabled online, and enabled a number of users etc. Everytime it seems like it comes out of some smaller software company with small bright teams. I'm guessing in these cases the senior level codes and maybe even the whole team got involved.
Anyone out there have expierence writing key checkers and other piracy related pieces of functionality?
is John Romero. Was he in that movie with that other guy?
Wow what a rant, that got up to 5 ( Insightful ). Way to go slashdot.
HTML/CSS/JavaScript like any technology is getting old. It wasn't designed to really be for applications. Now we have Ajax hacks and a slew of other crap to try and make it like a normal desktop app...things that flash and java applets ( yes I know applets are not that great ) just do.
Flash can be just as accesible if not more then a web page...it is all in the tools that make it accesible. Imagine if I wrote a flash app specifically for blind people...I'm guessing I could get a lot further then with just a web page.
Instead of trying to make a page accessible...i'd rather see a version of the app written specifically for blind people. It'd be better if google or other companies teamed up with another company, give them the raw content as XML and let them expose it in a way that will make it easier to access. Browser are inherently visual are they not....maybe it'd make more sense for google to try and expose the information in a way that could be converted to brail or audio easily.
Yes there are issues.
http://www.webaim.org/techniques/flash/
I'm sorry to the individuals out there that have disabilities. At the same time some content that is very hard if not impossible to make accessible can make it far easy to access for people without a disablity to use it. We need to find ways to appease both communities.
What does the poster mean by the "real Web". And they SHOULD just post PDFs on the side of their content, they are way more accesible then HTML. I mean I'm sure you could get a program to read the content of a PDF far easier then you could get it to strip out text from an html document and read it.
As much as I think this could help in the event of disaster, I hope that some legislation is passed to limit it's use. It would be very easy to abuse it for propogand purposes.
Linus stops patching Linux v 1.4
Funny that Web 2.0 is taking off so much. The problem with it is that everyone I interview is now "learning" Ajax. I feel like if I go to an interview I'll be asked a million Ajax questions, that I really don't want to answer.
:)
Using hidden Iframes and JScript was one way to do what Ajax does years ago. There are definately a few cases where it is really useful. A little div popup, pre-populating city state after a postal code was entered, testing a value etc. Debugging is much harder, and the Javascript/DOM model is hard to code bug free. Javascript errors don't get reported to the server admin, and they are often hard to replicate. This is partly a lack of good tools, but view source on HTML is almost always easier then trying to step thru some buggy jscript.
It can be very easy to abuse Ajax. I recently had someone show me a search example that "pre-populated" as you typed. It was super clunky and really didn't work. Ajax's biggest problem at this point is that everyone thinksd everything has to be instant now. You can make a user go to another page to edit something that is not edited every other minute.
As much as I love Google maps, Yahoo Flash maps kick their ass. Adobe's new Flex tech is really going to give Ajax a run for the money. Java is just to sluggish, but Flash is pretty quick. Yes you'll have to turn off your flash ad blockers.
The thing that has to happen is that SVG or a new standard needs to be born to handle GUI apps. People don't like flash because there is a name behind it, HTML is a standard, Javascript is a standard, etc. Java is Sun/IBM, Flash is Adobe ( formally Macromedia ).
Personally I would love to see an HTML 5.0....A pure XML based HTML is great, but pretty impractical given the huge amount of content that doesn't have the
tag, and just have
tags, etc. WTH did no one think to have a tag? Now I'm stuck with a million different Javascript/UL combos out there. Even adding a target to div would be great. Imagine a that would turn on a div and tell the browser to turn it on. With some style sheet properties you could make some powerful divs without code.
I guess my biggest gripe with Web 2.0 is that almost everything that we spend hours figuring out in JScript could be done if people would create more and better HTML tags. Then the browser developers take care of all the testing, and we will have more stable apps.
Personally I'm going down the Flash path. If you haven't tried Flex yet, labs.adobe.com, do yourself a favor and see what you've been missing....no I don't work for Adboe or even really like them
You can do more in less time, and you can create content that really looks good. I'd love to see a Flex slashdot version.
All of these ideas are old, and while high performing don't address the largest issue of all, cross kernel compatability.
Sure you can recompile and all that jaz, but I'd love to see a day where an app could run on any number of kernels out there. This creates real competetion.
What I'd like to see if a kernel more like a CPU. Instead of linking your kernel calls, you place them as if you where placing an Assembly call. Then we can have many companies and open source organizations writing versions of it.
As we move towards multi core cpus this could really lead to performance leads. Where one or more of many cores could be dedicated to the kernel operations listening for operations and taking care of them. No context switches needed, no privledge mode switching.
Drivers and everything else run outside of kernel mode and use low level microcode to execute the code.
The best part I think is you could make it backword compatiable as we re-write. A layer could handle old kernel calls and change them to the micro codes.
As we define everything more and more then we might even be able to design CPUs that can handle it better.