Well, then, thre's the start. Post your email address here, and a URL for the "These sites are broken under IE" web site you're planning to create (if you're still planning to do that) and I'll encourage all my friends to email you every time they come across a site they think is broken.
Ok, now your just being silly. There is no system in place. Merely a discussion so far. Go ahead and shoot me emails beforehand I'll get right on them.
A job gets done when it's placed in the hands of someone who is clued, motivated and authorized. Wiki's authorize everyone, use a community feedback mechanism to sort out the clued, and don't depend on the actions of anyone who is not self-motivated. The plan you are proposing depands on cooperation between (self-motivated and possibly clued, but not authorized) spotters, (self motivated and possibly clued, but not authorized) screeners, and (weakly motivated, clearly unclued, but fully authorized) web maintainers. No single party posseses all three components, so the job will likely never get done.
Authorization at any access level doesn't inherently limit a user to any particular group. ie. an admin could submit a report.
No, but you haven't thought this through. You've specified at least two layers of volunteers (spotters and screeners) who must collect the data, but cannot do anything with the data. The only ones who can do anything with the data are not volunteers.
I acutally specified at least 3 layers, don't forget collecting of email info:) There will also be the task of changing reports to fixed should a site be made to work correctly.
So let's walk this through. I'm browsing the web. I hit a site which doesn't render properly, and in order to get done what I need to get done I need it to render properly. I click my "report it" button and say "it doesn't render properly" and off goes the report. If everything works properly, a volunteer somewhere will screen the report and send a polite letter to the web administrator for the broken page asking him to fix the problem. But guess what? Any update to the page is weeks away and I still need to get done what I need to get done. After a few cycles of "click the button, fill-in the mandatory description, fail to see results" I'm going to start hesitating to click that button, because clicking that button means I need to fill-in the mandatory description, and I'm not browsing the web in order to fill-in a mandatory description, I'm browsing the web to get done what I need to get done. Soon after, the mandatory description will start getting nonsense values like "asdfjkl;" just so I can hit "submit" rather than "cancel" (which just wastes the screener's time) and at that point the plug-in might just as well uninstall itself.
Good point (the constructive parts of the criticism are indeed welcome). I guess it makes the statistics page all that much more inportant as immediate feedback to the submitter.
So, part of this Firefox plugin will launch IE in the background, so that people can compare the pages, or are you saying that people who stumble across a page which they suspect isn't rendering properly in Firefox must then launch IE manually to view the page, make the comparison, and file the report?
Ok, you're being silly again. Automatically launching IE is overkill and a bad idea to do from firefox anyway. No sense in spawning a child process when a user can verify something is IE only simply by launching IE and checking for themselves.
If a web page on my site is broken in that it contains a broken link, I'll see the report in/var/log/httpd/error_log as soon as someone clicks on it. I do review that log and fix anything reported through this process. A simple extention of this requiring no software development at all would have people discovering http://example.net/broken_page.html doesn't render properly in their browser to
Are you volunteering to do this filtering, or just pay others to filter?;-)
As I said before it would rely on many volunteers. I wouldn't mind volunteering, its not like its hard work, and it has the potential to help spread the use of firefox.
If it isn't automatic, it won't happen. Most people never even change the default security setting of their brower.
Why are wiki sites successfull? Peoeple seem willing enough to donate information if they like the cause. I for one would gladly click a button on my browser when I came accross something that didn't work in firefox.
How do I know if a site is broken, or if the webmaster is just being creative? Sometimes a site can be broken without leaving obviously visible artifacts. Other times the unreadable rendering results from incorrect client-side settings.
Is the website functional in IE, but not in firefox or vise versa? Does the website have a piece of javascript that barks at you to use a specific browser? None of this is rocket science.
I suspect you'd see a lot of false positives or incorrect usage comments such as "that color looks ugly" or "how do I get to your competitor's homepage".
Are you going to just ignore the fact that I already stated this project would require volunteers to do something with the data?
I think most people would leave the description field blank, or worse, be unable to provide a useful comment. How would the average person describe buttons which didn't render, if their browser gave no hint that there were supposed to be buttons in the first place?
I suppose you could make description a required field. I assume some comments would be of the variety: "firefox isn't showing the navigation menu on the site, but i shows up fine in IE"
If the sitemaster is ignoring logs, you've got no chance of changing their behavior with anything short of a lawsuit. You likely couldn't even get their attention by driving their traffic to zero, because they're ignoring their logs. If they're getting any sort of click-through revenus, though, they won't be ingoring their logs.
Possibly, but your proposal of changing the HTTP standard to allow for a browser to give a response header of whether or not it successfully rendered is at best an optimisitic outlook of changes several years from now. I'm proposing a system to try to help the situation as webservers and clients operate today.
There is no reliable way to associate an email address with a site. You could 'guess' something like "webmaster@domain" but anything that obvious is either already blackholed against spam, or would be as soon as you started sending 'friendly' complaints to that address. What about hosting companise? Who gets the report if http://sourceforge/myproject is broke?
Again, some human data mining will be necessary. WHOIS databases typically contain technical and administrative contact info. Not nearly 100% accurate, but a start on top of any links that may exist on the websites themselves.
Broken site or broken page? Would one page broken on a site make the whole site 'broken'? And who is running this highly rated site of pages you wouldn't bother visiting? More to the point, who is their lawyer, because if they say something negative like that about one of my pages, and it reduces my site traffic or revenue, I'm going to sue. After all, broken html is not a crime.
Gimme a break. A public database of cases of websites not working with certain browsers is hardly libel.
And if Microsoft were to adopt this, would it also list sites broken to IE? If they did, most of the reports would be from IE users, considering their market share. I pity the poor Mozilla foundation employee spending all his time politely asking IIS webmins to fix their site so that IE users have a more pleasant web surfing experience...
I don't think that would scale very well. I suspect the number of broken web page views that occur on the internet per day, even by Firefox users alone, would put the amount of spam received by an unfiltered account to shame.
This is where the human element comes in. Nothing should go through unfiltered. No resposnes should be sent out automatically. If you come across a broken site, you hit the feedback button and put in a short description. It could also send a small amount of data to the database including things like url(which would be filled in automatically),warped image text verification(to weed out bots),page title(automatic, possibly overrideable),browser version info.
Perhaps a better solution would be to encourage a web standard wherby the return code from a web browser rendering a web page could be sent back to the web server as part of an OPTIONS method call including the information about the URL being rendered.
That could be helpful, but it wouldn't sovle the entire problem. Someone with a broken website may be ignoring logs and need their feet held to the fire. Be it a friendly email or their website showing up as a high rated broken site on a hopefully popular statistics page.
Still full of holes (need to include state information, and possibly not only the browser info but information about plug-in versions as well) but at least it sends the necessary information to someone in a position to effect a change, rather than to some Mozilla foundation volunteer who really couldn't care less.
It doesn't necessarily need to be a current mozilla volunteer. It doesn't even need to be a mozilla project if it was just a popular plug-in. If it was something that provided just enough data to be useful and fix some websites that were IE only, it could be something that some people do in fact care about.
This may have been proposed before, but what if there was a standard way to deal with non standards compliant websites? What if there was a simple feedback form as part of firefox? These would send error reports to a database at mozilla or somewhere. The reports can be gone over and a standard polite email can be sent to the webmaster informing them of the problems with their websites.
There would be quite a bit work involved I imagine. Who collects webmaster email info? What would be a non-intrusive UI to handle this? Who would go over the database and remove troll submissions? What kind of system is used to validate submissions?
The thing is, there are tons of people that would like to help out an open source project like firefox but do not have much in the way of programming skill. Something like this maybe be a place where their volunteer ours could come in handy.
Power over ethernet from your laptop, google it. Adapters are cheap.
You're probably better with a power inverter into your lighter plug as you'll want to plug in your laptop if you're gonna drive around for a long time.
Fine, here's just a few:
Bridged mode for point to point. Think about extending two buildings as though an ethernet cable was simply connecting the two physical networks
Plain access point, not router
Promicuous mode for war driving
Mount to lan share to dump data for WEP cracking etc. etc.
I'm a software engineer not a network engineer but its easy enough to see the possibilities.
I bought mine in phoenix for 25 dollars a couple weeks ago. Brand new,not returned, not refurbished, and no rebate required. Fry's always has good sales on friday, I wouldn't be surprised if its 20 or 25 bucks tomorrow.
There's a somewhat active fork called Kompiere Libero that apparently has compiere working on postgres and they've added a manufacturing module Haven't tried it yet, but take a look
Compiere also says you're wrong. It is both an ERP and CRM and open source. Only drawback is that it currently requires an oracle database, but work is being done on a Postgresql port.
J2ME with a JNI wrapper? Really a question not a suggestion. I haven't done java on a pda since my palmIII was new:)
Jbuilder 2005 is great, I've been using it since its release. I'm slowly switching to eclipse now that the j2ee webtools has hit 0.7. The rich client platform is based on the 3.1 release as well, so they should be able to co-exist on one install.
I agree. Fedora splits up java stuff in different places and puts the java executable in/usr/bin.
To override this you need to put JAVA_HOME/bin before it in your PATH. If the whole thing was somewhere like in/usr/local/gpl_jdk then you could just have a symlink for it as/usr/local/java and also have things like/usr/local/jdk1.4 and/usr/local/jdk1.5 in there. PATH could then just include/usr/local/java/bin and you could easily switch JVMs just by changing the symlink.
I have found it to be too buggy to use for java at work, so I'm still doing most development using regular eclipse on top of the Sun jvm. Never had it crash the entire OS.
Only played a little with the c++ dev environment, so I can't vouch for stablility there. I am looking forward to trying out native eclipse to do some php development once phpeclipse.sourceforge.net upgrades to the 3.1 eclipse version from 3.0.x
Eclipse is indeed open source. In fact, fedora core 4 comes with a natively compiled version of eclipse and a 100% open source jvm implementation. Still needs quite a bit of work, but has definetly come a long way
Guess you are out of the loop:)
Take a look at the stats Actually, I only found out about the project a few months ago.
Check out the demo its actually a really cool editor. Amazing what this guy can do with javascript. Works great as an embedded editor for a web-based content management system.
Suse also does a good job of providing and configuring nvidia drivers. Not that its all that hard to get them installed with other distros, suse just makes it as easy as possible.
In the real world (yes, there are some exceptions) people use Microsoft Office.
Do you reall believe this tired "real world" argument? Does a high school diploma plus MS office on a resume land a good job? Will it in 10-15 years when grade school students enter the workforce? In the real world most users of office don't use any more than the functionality that is available for free in OpenOffice.
There is no point in teaching grade school students all the nuances of the latest Microsoft offering. Unless of course, teaching brand loyalty is part of the curriculum.
Take a look here Gives a rough estimate on sizing of the server.
In my experience, anything at least pentium1 class works fine for a client.
Also, if you want to get up and running quickly, be sure to pull down the K12LTSP distro. Its basically the latest fedora with ltsp pre-configured and a bunch of educational software thrown in.
Slashdot Mirror
Mambo is a really cool project and deserves an award. I just started using it a couple months ago and am really happy with it.
The article could have at least mentioned that LTSP took Best In Show overall.
Well, then, thre's the start. Post your email address here, and a URL for the "These sites are broken under IE" web site you're planning to create (if you're still planning to do that) and I'll encourage all my friends to email you every time they come across a site they think is broken.
:) There will also be the task of changing reports to fixed should a site be made to work correctly.
/var/log/httpd/error_log as soon as someone clicks on it. I do review that log and fix anything reported through this process. A simple extention of this requiring no software development at all would have people discovering http://example.net/broken_page.html doesn't render properly in their browser to
Ok, now your just being silly. There is no system in place. Merely a discussion so far. Go ahead and shoot me emails beforehand I'll get right on them.
A job gets done when it's placed in the hands of someone who is clued, motivated and authorized. Wiki's authorize everyone, use a community feedback mechanism to sort out the clued, and don't depend on the actions of anyone who is not self-motivated. The plan you are proposing depands on cooperation between (self-motivated and possibly clued, but not authorized) spotters, (self motivated and possibly clued, but not authorized) screeners, and (weakly motivated, clearly unclued, but fully authorized) web maintainers. No single party posseses all three components, so the job will likely never get done.
Authorization at any access level doesn't inherently limit a user to any particular group. ie. an admin could submit a report.
No, but you haven't thought this through. You've specified at least two layers of volunteers (spotters and screeners) who must collect the data, but cannot do anything with the data. The only ones who can do anything with the data are not volunteers.
I acutally specified at least 3 layers, don't forget collecting of email info
So let's walk this through. I'm browsing the web. I hit a site which doesn't render properly, and in order to get done what I need to get done I need it to render properly. I click my "report it" button and say "it doesn't render properly" and off goes the report. If everything works properly, a volunteer somewhere will screen the report and send a polite letter to the web administrator for the broken page asking him to fix the problem. But guess what? Any update to the page is weeks away and I still need to get done what I need to get done. After a few cycles of "click the button, fill-in the mandatory description, fail to see results" I'm going to start hesitating to click that button, because clicking that button means I need to fill-in the mandatory description, and I'm not browsing the web in order to fill-in a mandatory description, I'm browsing the web to get done what I need to get done. Soon after, the mandatory description will start getting nonsense values like "asdfjkl;" just so I can hit "submit" rather than "cancel" (which just wastes the screener's time) and at that point the plug-in might just as well uninstall itself.
Good point (the constructive parts of the criticism are indeed welcome). I guess it makes the statistics page all that much more inportant as immediate feedback to the submitter.
So, part of this Firefox plugin will launch IE in the background, so that people can compare the pages, or are you saying that people who stumble across a page which they suspect isn't rendering properly in Firefox must then launch IE manually to view the page, make the comparison, and file the report?
Ok, you're being silly again. Automatically launching IE is overkill and a bad idea to do from firefox anyway. No sense in spawning a child process when a user can verify something is IE only simply by launching IE and checking for themselves.
If a web page on my site is broken in that it contains a broken link, I'll see the report in
Are you volunteering to do this filtering, or just pay others to filter? ;-)
As I said before it would rely on many volunteers. I wouldn't mind volunteering, its not like its hard work, and it has the potential to help spread the use of firefox.
If it isn't automatic, it won't happen. Most people never even change the default security setting of their brower.
Why are wiki sites successfull?
Peoeple seem willing enough to donate information if they like the cause. I for one would gladly click a button on my browser when I came accross something that didn't work in firefox.
How do I know if a site is broken, or if the webmaster is just being creative? Sometimes a site can be broken without leaving obviously visible artifacts. Other times the unreadable rendering results from incorrect client-side settings.
Is the website functional in IE, but not in firefox or vise versa?
Does the website have a piece of javascript that barks at you to use a specific browser? None of this is rocket science.
I suspect you'd see a lot of false positives or incorrect usage comments such as "that color looks ugly" or "how do I get to your competitor's homepage".
Are you going to just ignore the fact that I already stated this project would require volunteers to do something with the data?
I think most people would leave the description field blank, or worse, be unable to provide a useful comment. How would the average person describe buttons which didn't render, if their browser gave no hint that there were supposed to be buttons in the first place?
I suppose you could make description a required field. I assume some comments would be of the variety: "firefox isn't showing the navigation menu on the site, but i shows up fine in IE"
If the sitemaster is ignoring logs, you've got no chance of changing their behavior with anything short of a lawsuit. You likely couldn't even get their attention by driving their traffic to zero, because they're ignoring their logs. If they're getting any sort of click-through revenus, though, they won't be ingoring their logs.
Possibly, but your proposal of changing the HTTP standard to allow for a browser to give a response header of whether or not it successfully rendered is at best an optimisitic outlook of changes several years from now.
I'm proposing a system to try to help the situation as webservers and clients operate today.
There is no reliable way to associate an email address with a site. You could 'guess' something like "webmaster@domain" but anything that obvious is either already blackholed against spam, or would be as soon as you started sending 'friendly' complaints to that address. What about hosting companise? Who gets the report if http://sourceforge/myproject is broke?
Again, some human data mining will be necessary. WHOIS databases typically contain technical and administrative contact info. Not nearly 100% accurate, but a start on top of any links that may exist on the websites themselves.
Broken site or broken page? Would one page broken on a site make the whole site 'broken'? And who is running this highly rated site of pages you wouldn't bother visiting? More to the point, who is their lawyer, because if they say something negative like that about one of my pages, and it reduces my site traffic or revenue, I'm going to sue. After all, broken html is not a crime.
Gimme a break. A public database of cases of websites not working with certain browsers is hardly libel.
And if Microsoft were to adopt this, would it also list sites broken to IE? If they did, most of the reports would be from IE users, considering their market share. I pity the poor Mozilla foundation employee spending all his time politely asking IIS webmins to fix their site so that IE users have a more pleasant web surfing experience...
Sounds li
I don't think that would scale very well. I suspect the number of broken web page views that occur on the internet per day, even by Firefox users alone, would put the amount of spam received by an unfiltered account to shame.
,warped image text verification(to weed out bots),page title(automatic, possibly overrideable),browser version info.
This is where the human element comes in. Nothing should go through unfiltered. No resposnes should be sent out automatically.
If you come across a broken site, you hit the feedback button and put in a short description. It could also send a small amount of data to the database including things like url(which would be filled in automatically)
Perhaps a better solution would be to encourage a web standard wherby the return code from a web browser rendering a web page could be sent back to the web server as part of an OPTIONS method call including the information about the URL being rendered.
That could be helpful, but it wouldn't sovle the entire problem. Someone with a broken website may be ignoring logs and need their feet held to the fire. Be it a friendly email or their website showing up as a high rated broken site on a hopefully popular statistics page.
Still full of holes (need to include state information, and possibly not only the browser info but information about plug-in versions as well) but at least it sends the necessary information to someone in a position to effect a change, rather than to some Mozilla foundation volunteer who really couldn't care less.
It doesn't necessarily need to be a current mozilla volunteer. It doesn't even need to be a mozilla project if it was just a popular plug-in. If it was something that provided just enough data to be useful and fix some websites that were IE only, it could be something that some people do in fact care about.
oops, hit the preview button next time:
Something like this may be be a place where their volunteer hours could come in handy.
You're probably right and it is unfortunate.
This may have been proposed before, but what if there was a standard way to deal with non standards compliant websites?
What if there was a simple feedback form as part of firefox? These would send error reports to a database at mozilla or somewhere. The reports can be gone over and a standard polite email can be sent to the webmaster informing them of the problems with their websites.
There would be quite a bit work involved I imagine. Who collects webmaster email info? What would be a non-intrusive UI to handle this? Who would go over the database and remove troll submissions? What kind of system is used to validate submissions?
The thing is, there are tons of people that would like to help out an open source project like firefox but do not have much in the way of programming skill. Something like this maybe be a place where their volunteer ours could come in handy.
Power over ethernet from your laptop, google it. Adapters are cheap.
You're probably better with a power inverter into your lighter plug as you'll want to plug in your laptop if you're gonna drive around for a long time.
Fine, here's just a few:
Bridged mode for point to point. Think about extending two buildings as though an ethernet cable was simply connecting the two physical networks
Plain access point, not router
Promicuous mode for war driving
Mount to lan share to dump data for WEP cracking
etc. etc.
I'm a software engineer not a network engineer but its easy enough to see the possibilities.
I bought mine in phoenix for 25 dollars a couple weeks ago. Brand new,not returned, not refurbished, and no rebate required.
Fry's always has good sales on friday, I wouldn't be surprised if its 20 or 25 bucks tomorrow.
If you could get the same functionality for less than half the price, wouldn't that be a good thing?
It will likely be used to extend functionality
The popular linksys G router has a linux firmware that people have done some really cool things with.
two weeks ago from fry's.
.
still in package. most definetly brings a smile.
However, its just a project that "aims" to devise improved Linux firmware
There's a somewhat active fork called Kompiere Libero that apparently has compiere working on postgres and they've added a manufacturing module
Haven't tried it yet, but take a look
Compiere also says you're wrong.
It is both an ERP and CRM and open source.
Only drawback is that it currently requires an oracle database, but work is being done on a Postgresql port.
J2ME with a JNI wrapper? Really a question not a suggestion. I haven't done java on a pda since my palmIII was new :)
Jbuilder 2005 is great, I've been using it since its release. I'm slowly switching to eclipse now that the j2ee webtools has hit 0.7. The rich client platform is based on the 3.1 release as well, so they should be able to co-exist on one install.
Exactly!
What we need is IBM's JVM to be open-sourced PLUS a finished GNU classpath.
I agree. Fedora splits up java stuff in different places and puts the java executable in /usr/bin.
To override this you need to put JAVA_HOME/bin before it in your PATH. /usr/local/gpl_jdk then you could just have a symlink for it as /usr/local/java and also have things like /usr/local/jdk1.4 and /usr/local/jdk1.5 in there. /usr/local/java/bin and you could easily switch JVMs just by changing the symlink.
If the whole thing was somewhere like in
PATH could then just include
Crashes the OS?
I have found it to be too buggy to use for java at work, so I'm still doing most development using regular eclipse on top of the Sun jvm.
Never had it crash the entire OS.
Only played a little with the c++ dev environment, so I can't vouch for stablility there.
I am looking forward to trying out native eclipse to do some php development once phpeclipse.sourceforge.net upgrades to the 3.1 eclipse version from 3.0.x
Eclipse is indeed open source.
In fact, fedora core 4 comes with a natively compiled version of eclipse and a 100% open source jvm implementation.
Still needs quite a bit of work, but has definetly come a long way
Guess you are out of the loop :)
Take a look at the stats Actually, I only found out about the project a few months ago.
Check out the demo its actually a really cool editor. Amazing what this guy can do with javascript.
Works great as an embedded editor for a web-based content management system.
Suse also does a good job of providing and configuring nvidia drivers.
Not that its all that hard to get them installed with other distros, suse just makes it as easy as possible.
In the real world (yes, there are some exceptions) people use Microsoft Office.
Do you reall believe this tired "real world" argument? Does a high school diploma plus MS office on a resume land a good job? Will it in 10-15 years when grade school students enter the workforce?
In the real world most users of office don't use any more than the functionality that is available for free in OpenOffice.
There is no point in teaching grade school students all the nuances of the latest Microsoft offering. Unless of course, teaching brand loyalty is part of the curriculum.
Take a look here
Gives a rough estimate on sizing of the server.
In my experience, anything at least pentium1 class works fine for a client.
Also, if you want to get up and running quickly, be sure to pull down the K12LTSP distro. Its basically the latest fedora with ltsp pre-configured and a bunch of educational software thrown in.
oops. you're rite :)
Google has put all its eggs, as it were, in a single basket
Where have been for the last several months?
Google has been buying everything in site. Spreading out quite a bit.