1) I'd sell advertising on the side 2) I'd charge $0.25 for 15 minutes (with no ability to add $0.25 from inside).
Seriously, if you godda drop a duce, are you gonna use a free bathroom, or are you gonna beg borrow and steat a quarter to go to a pay restroom? Now imagine your wife.
Keep in mind these are all psychological, not "real". People will perceive the pay-restroom to be higher quality, better maintained (even if it isn't), and more sanitary.
Adding advertising makes it blend in with the fabric of the street. Right now, those damn things look like space age robots--very imposing.
you know what... before you posted this I was gonna side with the "why ALT everything" crowd but now I think you are right. I always add blanks anyway to shut up my validator, but now I feel a bit less like a tool:-)
Visual Studio 2005's C# functionality has on-the-fly code editing for minor changes. Minor nitpick, but this is true only for 32-bit code--for both 2005 & 2008. Once you are in 64-bit land, it just gives you a fatty "sorry pal, no such luck" error instead.
The best improvement is that Visual Studio now uses your XML documentation for a function while you type a call to it. 2005 was kind of flaky with those and maybe 50% of the time did it show them while you were typing in a function call. In fact, I'd go so far as to say the improvements to the XML documentation system are one of the best improvements they made. When you type "///", visual studio will automatically insert the entire XML template so you can fill it in. In addition, it green underlines any function missing the documentation.
/// <summary> /// This function is the best operator ever /// </summary> /// <param name="dim">Give me your dimension</param> /// <returns>Me!</returns> static public explicit operator AspectRatio(Dimension dim) { /* my code goes here */
The MySQL guys are like 75% of the way there (they already have pretty good.NET / Visual Studio integration). It is my beloved PostgreSQL that really lacks good Visual Studio (and by proxy, LINQ) integration. There is some dude who is pushing out a Linq provider for MySQL, Postgres and Oracle. Looks like there has been an update just a couple days ago even.
I dont blame anybody for why it is taking so long. I tried to get get my hands wet in the Linq provider stuff, and man is it tedious!
Data diodes are so old school man! It's all fiber-in-the-cabin now days. Clearly they need some kind of fiber optic polarizer instead to keep the riff raft from tapping into the important bits.
I think Office 2007 is a huge step forward, especially for Word. The designers took a rats nest of features and laid them all out in a clean, easy to discover way. Better still, it took advantage of modern hardware so you can preview what a given tool will do to your work before committing yourself to it.
The real crazy thing about it is that prior to Office 2007, people always bitched about how bloated Office was. To a point, I agreed. With Office 2007, I almost feel like Word doesn't have enough stuff in it! Everything it can possibly do in Word right in front of me now!
My question is, can this design ever work in an application like Visual Studio? What about Photoshop?
There exists no book that will teach you good user interface design. Buying a book about "user interface design" is like buying a book on "Learn Visual Basic in 24 hours"--too sepecific. You want the usability version of "code complete" instead--one that is all about theory and "why".
The most incredible and exciting thing about good design is it involves lots of *watching* and *listening* to the people who will use your software. There are no rules of interface design besides "dont violate the costraints of your medium (i.e. dont roll your own widget set"). Like "the rule of thirds" in art, those too are only soft rules and can be broken if you know what you are doing.
Good usability comes from good requirements gathering. You have to get out of your office and *listen* and *watch* the people who will use your work.
The book should spend very, very, very little time on heuristics and "rules of good interface design" because quite frankly, there are very few rules. The best book will probably not have screen shots of stupid error messages; it should be mostly wireframes and paper prototypes. The best book will only have about 1 sentence about fitts law. "Rules" are obvious; users are not.
You want a book that covers how to interview. The book should cover how to write a good persona. The book should talk about how to *correctly* screen for participants (like why you always ask the risky questions like somebodies income last and why you bracket it). The book should talk about how to *correct* do a usability test so you get valid results (for example, never try to correct "mistakes" or help them if they get lost... instead always ask what they expected to find and than shut up and watch). The book should dedicate an entire chapter to cardsorting and why it is so damn useful. The book should dedicate at least half a chapter to paper prototypes.
You do not want a book on "how to write a good interface". You need theory and practice. Usability is the most fun thing you can do in computing. *Nothing* is more exciting than spending two hours watching how somebody works and thinking of about a billion ways you can improve their experience in a way they would appreciate.
Re:Portsentry a good idea?
on
Linux Firewalls
·
· Score: 2, Insightful
Actually, it is the first consideration I have. I don't use software whose development seems to be dead. The first thing I look at on a website is "Last Updated $NOW - (ONE YEAR)". If it hasn't been touched in a year, I keep right on movin'...
In theory, you are correct. However, a fair number of evil software in the wild gets into your computers blood stream by you voluntarily running it. You know there are trojans that imitate common spyware removal programs? What about some "download accelerator" or "Pimp my Firefox Toolbar v3.423"?
It cant automatically run its self. Sure it can. Find me a box running PHP and some one year old web application written on top of it and I'll execute it for you!:-)
I'm not saying ACL's are bad. I think ACL's would be a very useful addition to Linux. Owner,Group,World is just not granular enough for big groups of people. I dont use ACL's on linux because they are just to much of a pain in the ass. I even turn them off on the filesystem so if somebody hacks the box they cannot bury shit in the ACL structure in a way I cannot see it using ls -l (and yes, I know it gives a hint that there are ACL's on a file).
I'm saying in their current state, ACL's are unusable mainly because the tool support for them sucks hard rocks. It is really hard to pull off a good ACL implementation using a 100% green screen command line interface. You need at least color, and even better a rich GUI.
Wow... late to the party?
The BSD license is more lax than the GPL, but they're similar in spirit Not even close. Not even in the same ballpark or even state.
FreeBSD is not boring Of course not my friend. It is very well organized and extremely well documented. All kinds of cool stuff come out and all of it is very exciting. But when it actually is running on a production server doing useful things, it should be boring. You don't want "excitement" on production servers, do you?
This is largely a cultural thing left over from unix, which discarded ACLs in order to save space on the file system Besides being a cultural thing, traditional unix filesystem permissions are far to entrenched for ACL's to take over. Does tar preserve anything but "owner, group, world"? Does Apache cope with ACL's? PHP? In all my travels, I've never once used ACLs or seen them used. Has anybody used them?
The bigger problem with ACL's in unix is the command line isn't a very good way to handle complex permissions. Complex permissions require complex displays. For example, using nothing but text, how do you show that for a given directory, a role gets it's permissions via inheritance, or by being explicitly set for that directory? In a GUI, you'd use a tri-state check box and make them "grey checked" if they inherited the permission. You could do it even in curses. The only place it gets tricky is trying to do it straight on the command line, especially if you say "no color allowed".
Really, the only way for ACL's to catch on is better tools. The only "better tools" are either straight GUI's or something using curses.
I know some people might disagree with me, but I'm beginning to think some open source projects would benefit from using a year for the public version number:
While internally, the product could use the same version scheme it did before, I think many open source projects are far too anal about version numbers. The stubborn refusal to bump up the "big" version field doesn't help public image because if it never moves up people think the project is dead.
The only version number that matters is the build number and repository version, the rest is marketing. Granted the year scheme isn't perfect in the early stages of a product when functionality is drastically changing every 3 months, but on mature products, I think we could all really benefit from number schemes that use the year the product was release.
If you only have 50 purchases a year, it means you probably know each person, right? As long as you know the process well enough you just walk them through it.
Adding you as a certificate authority doesn't scale, obviously. But I'll assert that if you get to a level where walking people through adding you is a hassle, you should be doing enough business that getting a code cert isn't a big deal. $250 per year isn't that big of a deal. It is just another cost of doing business. If you are doing 50 customers a year, that adds $5 per unit. How much are each of the units sold for?
You can run all your ancient 16-bit HVAC software in a virtual machine. I say HVAC because that whole industry still seems to think we all have a copy of MSDOS for their programs. Solution? XP and VMWare Server hosting a real copy of MSDOS 6.2 from MSDN. Since you dont have a floppy, you have to mount the floppy disk image and boot from that. The only trick is getting the network stack running. Remember autoexec.bat and config.sys?
Those old geezers really should be running in a virtual machine anyway. They tend to get confused when you run them on new hardware. Nothing is better than when they use signed 32-bit ints for disk space values and reporting you have -2147483647mb of disk space.
Speak for yourself. It took three hours to get the thing to work with my Phillips socket. IRC was no help. Check it out: me> Can somebody tell me how to get Linux working with my apartments Phillips brand lightbulb sockets? linuxd0rk> STFU n00bie! Everybody knows Phillips sucks! They dont want to support Linux Lightbulbs! me> That is nice, but I just signed a year lease for this aparment, what do you suggest? linuxd0rk> You suck! You should have looked at the lightbulb sockets before signing the lease. Everybody knows that! me> Is there a list of supported sockets? I tried to find one, but I couldn't. linuxd0rk> Google it, dumbass me> I tried google, no dice linuxd0rk> There is the linux hardware list, stupid. What, are you a westinghouse looser? me> That hardware is five years old, they dont even have the new sockets from Westinghouse that use AC power! linuxd0rk> Everybody knows AC is just an attempt by westinghouse further their monopoly. You should use DC, stupid. AC is a patent minefield and besides DC is way better anyways. It is open source power! me> DC? My apartment runs on 120V AC like everybody else in this country. What do you suggest? linuxd0rk> You should have got a DC apartment. DC is open source. Are you a westinghouse shill? Do you like all the evil patents on AC? me> AC is the defacto standard and has been for years, idiot! * me@123.abc.street.com banned by linuxd0rk (+xyz) ("STFU n00bie") bsdsucks> lol! stupid n00bies linuxd0rk> yeah. stupid n00bies... bsdsucks> wonder how much that guy got paid to shill out AC current? linuxd0rk> yeah. lots of westinghouse shills on this channel now...
Wow... how many hours did you spend doing this? If it was more than a couple hours, you probably spent more billable time dicking around than just ordering another stick of ram. Once you start dicking with services, odds are good you might make things more brittle too.
Time is money. In most cases, hardware is a lot cheaper than labor.
The folk behind FreeBSD are all just professional people trying to get real work done. You will not find your next religion when you use FreeBSD. Nor will you find saints who preach to you about how you should and should not use your computer. Nor will you find people telling you how you should and should not value your labor.
We dont give a shit if you like Vista. We dont care make FreeBSD work with your Windows Server. We dont care if you embed FreeBSD in your Tivo or Playstation. We are more than happy if you take our code and use it in your TCP/IP stack. Seriously. Take our code! No strings attached!
And hey, we all have to eat here in FreeBSD-land and so do you! We don't care if you make money from what comes out of your brain. Many of us are programmers whose livelihood depends on selling the value of our brain. We dont preach to you about the evils of intellectual property. If you sell software, the more power to you! If you become the next Microsoft, sweet!
And at the end of the day, FreeBSD works. It is the most boring OS you'll ever find. It is about as exciting as your water heater. And that is the best part about it.
I see what you are asking. Okay... think like this:
1) You are writing an application that will run on unix. This means make;./configure; make install. This means writing code that doesn't assume a specific compiler (people targeting linux are notorious for assuming gcc exists). This means you are distributing a source code tarball that looks and works just like all the other source code tarballs you are used to working with. 2) You happen to be offering it to FreeBSD users with a port. Ports are nothing but scripts that know how to interact with your make file. They are makefiles that script how to interact with a package's makefiles. The port makefile says "here is where to download the source code, here is how to unpack it, here is a list of everything that gets installed, here is how to call./configure, here are the arguments for./configure, here are some patches, and here is where I want you to install it"
Take a look at the ports tree. There are some doozies in there like apache, perl or php so avoid them at first. Look at how some silly app is installed via a port and copy it for now. If you are writing a language library (like a CPAN module) make sure to see how others install CPAN modules, same with PEAR modules.
Make sure to read the porters handbook too! On FreeBSD, there is actually good documentation to work with so enjoy your new OS:-)
Remember, you aren't writing a program for freebsd. You are writing a program for unix, and you are writing a wrapper to install it on freebsd.
Slashdot Mirror
Or is the design different?
Do they sell advertising on the side?
Do they charge for getting in?
Do they look more like eurotrash artwork than a bathroom?
If I designed them, I'd do two things:
1) I'd sell advertising on the side
2) I'd charge $0.25 for 15 minutes (with no ability to add $0.25 from inside).
Seriously, if you godda drop a duce, are you gonna use a free bathroom, or are you gonna beg borrow and steat a quarter to go to a pay restroom? Now imagine your wife.
Keep in mind these are all psychological, not "real". People will perceive the pay-restroom to be higher quality, better maintained (even if it isn't), and more sanitary.
Adding advertising makes it blend in with the fabric of the street. Right now, those damn things look like space age robots--very imposing.
you know what... before you posted this I was gonna side with the "why ALT everything" crowd but now I think you are right. I always add blanks anyway to shut up my validator, but now I feel a bit less like a tool :-)
Yet.
.NET / Visual Studio integration). It is my beloved PostgreSQL that really lacks good Visual Studio (and by proxy, LINQ) integration. There is some dude who is pushing out a Linq provider for MySQL, Postgres and Oracle. Looks like there has been an update just a couple days ago even.
The MySQL guys are like 75% of the way there (they already have pretty good
I dont blame anybody for why it is taking so long. I tried to get get my hands wet in the Linq provider stuff, and man is it tedious!
I'm late to the party.
Data diodes are so old school man! It's all fiber-in-the-cabin now days. Clearly they need some kind of fiber optic polarizer instead to keep the riff raft from tapping into the important bits.
I think Office 2007 is a huge step forward, especially for Word. The designers took a rats nest of features and laid them all out in a clean, easy to discover way. Better still, it took advantage of modern hardware so you can preview what a given tool will do to your work before committing yourself to it.
The real crazy thing about it is that prior to Office 2007, people always bitched about how bloated Office was. To a point, I agreed. With Office 2007, I almost feel like Word doesn't have enough stuff in it! Everything it can possibly do in Word right in front of me now!
My question is, can this design ever work in an application like Visual Studio? What about Photoshop?
There exists no book that will teach you good user interface design. Buying a book about "user interface design" is like buying a book on "Learn Visual Basic in 24 hours"--too sepecific. You want the usability version of "code complete" instead--one that is all about theory and "why".
The most incredible and exciting thing about good design is it involves lots of *watching* and *listening* to the people who will use your software. There are no rules of interface design besides "dont violate the costraints of your medium (i.e. dont roll your own widget set"). Like "the rule of thirds" in art, those too are only soft rules and can be broken if you know what you are doing.
Good usability comes from good requirements gathering. You have to get out of your office and *listen* and *watch* the people who will use your work.
The book should spend very, very, very little time on heuristics and "rules of good interface design" because quite frankly, there are very few rules. The best book will probably not have screen shots of stupid error messages; it should be mostly wireframes and paper prototypes. The best book will only have about 1 sentence about fitts law. "Rules" are obvious; users are not.
You want a book that covers how to interview. The book should cover how to write a good persona. The book should talk about how to *correctly* screen for participants (like why you always ask the risky questions like somebodies income last and why you bracket it). The book should talk about how to *correct* do a usability test so you get valid results (for example, never try to correct "mistakes" or help them if they get lost... instead always ask what they expected to find and than shut up and watch). The book should dedicate an entire chapter to cardsorting and why it is so damn useful. The book should dedicate at least half a chapter to paper prototypes.
You do not want a book on "how to write a good interface". You need theory and practice. Usability is the most fun thing you can do in computing. *Nothing* is more exciting than spending two hours watching how somebody works and thinking of about a billion ways you can improve their experience in a way they would appreciate.
Actually, it is the first consideration I have. I don't use software whose development seems to be dead. The first thing I look at on a website is "Last Updated $NOW - (ONE YEAR)". If it hasn't been touched in a year, I keep right on movin'...
I'm not saying ACL's are bad. I think ACL's would be a very useful addition to Linux. Owner,Group,World is just not granular enough for big groups of people. I dont use ACL's on linux because they are just to much of a pain in the ass. I even turn them off on the filesystem so if somebody hacks the box they cannot bury shit in the ACL structure in a way I cannot see it using ls -l (and yes, I know it gives a hint that there are ACL's on a file).
I'm saying in their current state, ACL's are unusable mainly because the tool support for them sucks hard rocks. It is really hard to pull off a good ACL implementation using a 100% green screen command line interface. You need at least color, and even better a rich GUI.
Even still you can "double click to install RPM", and that is just as good as an executable...
The bigger problem with ACL's in unix is the command line isn't a very good way to handle complex permissions. Complex permissions require complex displays. For example, using nothing but text, how do you show that for a given directory, a role gets it's permissions via inheritance, or by being explicitly set for that directory? In a GUI, you'd use a tri-state check box and make them "grey checked" if they inherited the permission. You could do it even in curses. The only place it gets tricky is trying to do it straight on the command line, especially if you say "no color allowed".
Really, the only way for ACL's to catch on is better tools. The only "better tools" are either straight GUI's or something using curses.
But it was under development the whole time.
I know some people might disagree with me, but I'm beginning to think some open source projects would benefit from using a year for the public version number:
Octave 2008 (3.0.x)
Thunderbird 2006 (2.0.x)
Firefox 2008 (3.0.x)
FreeBSD 2006 (6.0)
FreeBSD 2008 (7.0)
PostgreSQL 2006 (8.1)
PostgreSQL 2007 (8.2)
PostgreSQL 2008 (8.3)
While internally, the product could use the same version scheme it did before, I think many open source projects are far too anal about version numbers. The stubborn refusal to bump up the "big" version field doesn't help public image because if it never moves up people think the project is dead.
The only version number that matters is the build number and repository version, the rest is marketing. Granted the year scheme isn't perfect in the early stages of a product when functionality is drastically changing every 3 months, but on mature products, I think we could all really benefit from number schemes that use the year the product was release.
If you only have 50 purchases a year, it means you probably know each person, right? As long as you know the process well enough you just walk them through it.
Adding you as a certificate authority doesn't scale, obviously. But I'll assert that if you get to a level where walking people through adding you is a hassle, you should be doing enough business that getting a code cert isn't a big deal. $250 per year isn't that big of a deal. It is just another cost of doing business. If you are doing 50 customers a year, that adds $5 per unit. How much are each of the units sold for?
You can run all your ancient 16-bit HVAC software in a virtual machine. I say HVAC because that whole industry still seems to think we all have a copy of MSDOS for their programs. Solution? XP and VMWare Server hosting a real copy of MSDOS 6.2 from MSDN. Since you dont have a floppy, you have to mount the floppy disk image and boot from that. The only trick is getting the network stack running. Remember autoexec.bat and config.sys?
Those old geezers really should be running in a virtual machine anyway. They tend to get confused when you run them on new hardware. Nothing is better than when they use signed 32-bit ints for disk space values and reporting you have -2147483647mb of disk space.
Speak for yourself. It took three hours to get the thing to work with my Phillips socket. IRC was no help. Check it out:
me> Can somebody tell me how to get Linux working with my apartments Phillips brand lightbulb sockets?
linuxd0rk> STFU n00bie! Everybody knows Phillips sucks! They dont want to support Linux Lightbulbs!
me> That is nice, but I just signed a year lease for this aparment, what do you suggest?
linuxd0rk> You suck! You should have looked at the lightbulb sockets before signing the lease. Everybody knows that!
me> Is there a list of supported sockets? I tried to find one, but I couldn't.
linuxd0rk> Google it, dumbass
me> I tried google, no dice
linuxd0rk> There is the linux hardware list, stupid. What, are you a westinghouse looser?
me> That hardware is five years old, they dont even have the new sockets from Westinghouse that use AC power!
linuxd0rk> Everybody knows AC is just an attempt by westinghouse further their monopoly. You should use DC, stupid. AC is a patent minefield and besides DC is way better anyways. It is open source power!
me> DC? My apartment runs on 120V AC like everybody else in this country. What do you suggest?
linuxd0rk> You should have got a DC apartment. DC is open source. Are you a westinghouse shill? Do you like all the evil patents on AC?
me> AC is the defacto standard and has been for years, idiot!
* me@123.abc.street.com banned by linuxd0rk (+xyz) ("STFU n00bie")
bsdsucks> lol! stupid n00bies
linuxd0rk> yeah. stupid n00bies...
bsdsucks> wonder how much that guy got paid to shill out AC current?
linuxd0rk> yeah. lots of westinghouse shills on this channel now...
Proof that linux isn't ready for the apartment!
25 watts is good enough for anybody!
You again!
Did you ever see my reply to the last discussion we had about signed drivers?
You can get Vista to trust your own private certificate authority. Check out my comments above and see if any of them apply to signing kernel drivers.
Let me know if any of that helps!!
Wow... how many hours did you spend doing this? If it was more than a couple hours, you probably spent more billable time dicking around than just ordering another stick of ram. Once you start dicking with services, odds are good you might make things more brittle too.
Time is money. In most cases, hardware is a lot cheaper than labor.
Have the brakelight be a circle of LED rings. Light braking = only the outer ring light. Light the inner rings up as a function of braking intensity.
You'll love it here. We are chill.
The folk behind FreeBSD are all just professional people trying to get real work done. You will not find your next religion when you use FreeBSD. Nor will you find saints who preach to you about how you should and should not use your computer. Nor will you find people telling you how you should and should not value your labor.
We dont give a shit if you like Vista. We dont care make FreeBSD work with your Windows Server. We dont care if you embed FreeBSD in your Tivo or Playstation. We are more than happy if you take our code and use it in your TCP/IP stack. Seriously. Take our code! No strings attached!
And hey, we all have to eat here in FreeBSD-land and so do you! We don't care if you make money from what comes out of your brain. Many of us are programmers whose livelihood depends on selling the value of our brain. We dont preach to you about the evils of intellectual property. If you sell software, the more power to you! If you become the next Microsoft, sweet!
And at the end of the day, FreeBSD works. It is the most boring OS you'll ever find. It is about as exciting as your water heater. And that is the best part about it.
I see what you are asking. Okay... think like this:
./configure; make install. This means writing code that doesn't assume a specific compiler (people targeting linux are notorious for assuming gcc exists). This means you are distributing a source code tarball that looks and works just like all the other source code tarballs you are used to working with. ./configure, here are the arguments for ./configure, here are some patches, and here is where I want you to install it"
:-)
1) You are writing an application that will run on unix. This means make;
2) You happen to be offering it to FreeBSD users with a port. Ports are nothing but scripts that know how to interact with your make file. They are makefiles that script how to interact with a package's makefiles. The port makefile says "here is where to download the source code, here is how to unpack it, here is a list of everything that gets installed, here is how to call
Take a look at the ports tree. There are some doozies in there like apache, perl or php so avoid them at first. Look at how some silly app is installed via a port and copy it for now. If you are writing a language library (like a CPAN module) make sure to see how others install CPAN modules, same with PEAR modules.
Make sure to read the porters handbook too! On FreeBSD, there is actually good documentation to work with so enjoy your new OS
Remember, you aren't writing a program for freebsd. You are writing a program for unix, and you are writing a wrapper to install it on freebsd.
Good luck!