Oh I can pick up a new language no problem. And mostly by looking at examples, and then modifying them to my own use.
BUT,
While you poo-poo the details, HOW you use that language makes a big difference. Does it have pointers? Well then you can use them. What, no pointers? Ok, then a different way of organizing information. Structures, types, classes, associative arrays? Hmmm?
Each way requires a different mind-set to effectively use the language. The difference between a new comer and someone who has used a language for a while is painfully clear.
BTW, I am mostly self taught, though I score over 95% on ANY language or CS course I have ever taken, be it night school or full length university level program. I suppose that you will use the "self-taught" as a detriment, and 25 or so years experience does not count....
If syntax is so much of a giant hurdle to get into a new language, then perhaps you should choose another field.
A good CS major should be able to pick up any language because the concepts are mostly the same.
Just like a "people" manager should be able to manage any given group of people?
The devil is in the details. Each language has its own quirks, which you MUST know if you are to be effective in it. I have had training in over 15 languages, and am current in three. Yet each new language still requires a learning curve.
Comma here, brace there, tildes, dollar signs, asterixes, brackets, ahhh it makes my head hurt sometimes....
Ok, maybe not for very well defined specs, for minor complexity.
But anything which is complex and/or has inter-dependancies, and you really do need to hold face to face meetings. Nothing beats a good white-boarding session. Which is why out-sourcing development from specs is running into problems.
Thie is the downside to having a huge, inter-dependent set of apps. and thousands of coders working on largely interdependent projects and That's why OSS methodologies have a bit of an edge in this context
Ok, so having a large number of OSS developers, who are probably not even in the same country, working on projects is better?
*nix uses a "large number of small apps" paradigm. Which is why the pipe character is used so much in scripts. This makes each app(script) in the pipe chain vulnerable to changes in another app(script).
For instance if a script relies on a date beinbg returned by another app as m/d/y, and the date format changes to y.m.d, then the script breaks.
something that worked in most available browsers, I think I'd see a better compatibility mix from an Ajax / Javascript environment.
Huh?
Using JavaScript you are at the mercy of the Web browser and its implementation. At the most basic AJAX level, IE and Mozilla/Firefox use different methods to instantiate the object which actually does the background communication!
All MS needs to do is change their method, and now you are re-writing all your client side apps to test for yet another method.
Not that I am not using this, but it is only for light interaction.
Re:Good read so far
on
Ajax in Action
·
· Score: 2, Insightful
The central idea behind Ajax is pretty good IMHO- moving as much of the presentation-tier processing as possible to the client system.
Which to me sounds like a stab at re-inventing Java applets.
All the problems you describe (name spaces, libraries, etc) are already solved in Java. In addition, you are not at the mercy of browser JavaScript bugs.
The only downside is the initial startup time for the Java applet, as the local JVM must be loaded, THEN the applet.
As for the JVM version, you can check for this in your applet before you start the dependant code, and you can ask the user to d/w the JVM.
The technology Java Web Start provides an easy to use framework.
The problem with having a common development server is that if you code something that causes a crash, then you take down the common server. Running a personal server allows these little gaffes to be fixed without affecting other people.
Performance is not really an issue (just need RAM, lots of RAM). If a query takes 1-2 seconds, so what? One user, one query. I use an old 400MHz machine, which was surplus. It has 1G of RAM, but only a 20G hard drive. It runs both the application code (Tomcat/Java/JSP) and the database server (MS SQL).
Now the test server and definately the production server need to be up to spec, as they will need to handle load.
BTW, this setup processes pages (from first servlet start to the final JSP) in under 0.3 seconds. The test server does the same page in under 0.05 seconds. The production server is under the millisecond range and does not show a time except under heavy load.
1.a Get a test application server and production application server. Yes, you need both. The development server would be the developers workstation. 6.a Formalize the testing process by people other than the original developers. 6.b Write test cases. 6.c Do regression testing. Especially for "transperant to the user" changes. 8.a Have a naming standard for table columns. This will make your life SOOOO much easier. 11.a Where you do need hardcoded directories, externalize the locations in configuration files.
This is where meta-data and a database comes in. You need a large number of fields to describe what you are saving, then you can organize it (report) as you want with database queries.
Of course you need to actually enter all that information, but hey, I do NOT do content, that is the customer's problem...
Everyone is able to stop paying them whenever they want. You just have to do this funny thing called "follow the law."
Correct.
Except that most speed limits are set too low. It is assumed that most people will speed about 10/$units over the speed limit. Plus, they are set low for revenue generation.
With this system you will get a lot of slow drivers, increased time to get anywhere, and lots of frustrated drivers. And hopefully political pressure to raise limits.
I sometimes get to work at 140/km in a 100/km zone. Usually in a train of cars doing the same thing. The conditions make it safe: little traffic, lots of space between cars, dry roads, daylight.
I do research on virtual groups and social identity. Our names matter when we interact fora long time online. A name represents a person exclusively online.
Hmmmm, I set up a unique identity where I can. I must have over 20 different online personas.
So where I am RetroGeek here, I am something else on another board or newsgroup server.
Psych research in onthis question is showing the importance of our online names both to us as individuals and to the smooth running of social structures.
Ah, but I can be whoever I want to be. All I need to do is create another email address and register.
I find it amusing that there are so many Web services which insist upon you telling them a valid email address to complete some registration process. I set up an email address, give it to the service, they email me a registration code, I delete the email address. Mozilla (SeaMonkey?) saves the ui/pw for that site, so I do not need to remember them.
While I wouldn't call maintenance coding simple, you seem to think it is much harder than it really is.
The real problem is getting all the interactions right. You tend to spend a lot of time finding out what code depends on what you are changing. Maybe that other code uses a workaround for buggy code you just fixed, which your fix will now break.
It is not the actual finding of bugs. It is tracing down all the interactions, and testing, testing, testing, testing.
The other problem is in creating a fork which only YOUR apps can use. So when the source owner (OSS or not) releases a fix, you have a problem on how to integrate it.....
Slashdot Mirror
Um, I hate to argue with you, but it was Kirk
Kirk said this, not Spock
Oh I can pick up a new language no problem. And mostly by looking at examples, and then modifying them to my own use.
BUT,
While you poo-poo the details, HOW you use that language makes a big difference. Does it have pointers? Well then you can use them. What, no pointers? Ok, then a different way of organizing information. Structures, types, classes, associative arrays? Hmmm?
Each way requires a different mind-set to effectively use the language. The difference between a new comer and someone who has used a language for a while is painfully clear.
BTW, I am mostly self taught, though I score over 95% on ANY language or CS course I have ever taken, be it night school or full length university level program. I suppose that you will use the "self-taught" as a detriment, and 25 or so years experience does not count....
If syntax is so much of a giant hurdle to get into a new language, then perhaps you should choose another field.
cute
A good CS major should be able to pick up any language because the concepts are mostly the same.
Just like a "people" manager should be able to manage any given group of people?
The devil is in the details. Each language has its own quirks, which you MUST know if you are to be effective in it. I have had training in over 15 languages, and am current in three. Yet each new language still requires a learning curve.
Comma here, brace there, tildes, dollar signs, asterixes, brackets, ahhh it makes my head hurt sometimes....
The location of the coders is irrelevant.
Oh yes it is relevant.
Ok, maybe not for very well defined specs, for minor complexity.
But anything which is complex and/or has inter-dependancies, and you really do need to hold face to face meetings. Nothing beats a good white-boarding session. Which is why out-sourcing development from specs is running into problems.
Thie is the downside to having a huge, inter-dependent set of apps.
and
thousands of coders working on largely interdependent projects
and
That's why OSS methodologies have a bit of an edge in this context
Ok, so having a large number of OSS developers, who are probably not even in the same country, working on projects is better?
*nix uses a "large number of small apps" paradigm. Which is why the pipe character is used so much in scripts. This makes each app(script) in the pipe chain vulnerable to changes in another app(script).
For instance if a script relies on a date beinbg returned by another app as m/d/y, and the date format changes to y.m.d, then the script breaks.
I need to get out more.
I actually understood this thread.....
something that worked in most available browsers, I think I'd see a better compatibility mix from an Ajax / Javascript environment.
Huh?
Using JavaScript you are at the mercy of the Web browser and its implementation. At the most basic AJAX level, IE and Mozilla/Firefox use different methods to instantiate the object which actually does the background communication!
All MS needs to do is change their method, and now you are re-writing all your client side apps to test for yet another method.
Not that I am not using this, but it is only for light interaction.
The central idea behind Ajax is pretty good IMHO- moving as much of the presentation-tier processing as possible to the client system.
Which to me sounds like a stab at re-inventing Java applets.
All the problems you describe (name spaces, libraries, etc) are already solved in Java. In addition, you are not at the mercy of browser JavaScript bugs.
The only downside is the initial startup time for the Java applet, as the local JVM must be loaded, THEN the applet.
As for the JVM version, you can check for this in your applet before you start the dependant code, and you can ask the user to d/w the JVM.
The technology Java Web Start provides an easy to use framework.
still hasn't come up with a solution that works for software
Ever heard of a dongle?
Yes, it is not a purely s/w solution, but it does prevent use at more than one location at a time.
Could somebody please tell me why people use PDF's in the first place?
If I send a Word document, then your version of Word may helpfully re-format it, if you can read it at all. Also, you can alter the contents.
I can generate a PDF. Now the formatting is the same on both computers, AND you cannot alter it (easily).
And I use WordPerfect, which has "Print to PDF" built in.
Wow. I'm pretty old, but I have no recollection of 12 ft. discs.
You should see my collection. I have a whole barn dedicated to storing them.
AND, I managed to find a jukebox to play them. It uses industrial strength hydraulics to change the dics.
because a democratic system means it's a popularity contest, and we all know popularity doesn't mean quality
But wait, this is how we elect politicians!
LOL
Sigh.
Serious should be seriously
the Page should be then Page
Yes, except the pagination system is serious screwed up.
Let's say that the system allows x posts per page. If the first comment has more than x replies, then you will NEVER see x+1, x+2, and so on replies.
Clicking on Page 2 simply retrieves Page 1. If there are enough replies, the Page 2, 3, etc also only retrieve Page 1.
Eventually you CAN see the next root comment, but all those first comment replies are lost to viewing.
Unless there is some option I have not found.....
The problem with having a common development server is that if you code something that causes a crash, then you take down the common server. Running a personal server allows these little gaffes to be fixed without affecting other people.
Performance is not really an issue (just need RAM, lots of RAM). If a query takes 1-2 seconds, so what? One user, one query. I use an old 400MHz machine, which was surplus. It has 1G of RAM, but only a 20G hard drive. It runs both the application code (Tomcat/Java/JSP) and the database server (MS SQL).
Now the test server and definately the production server need to be up to spec, as they will need to handle load.
BTW, this setup processes pages (from first servlet start to the final JSP) in under 0.3 seconds. The test server does the same page in under 0.05 seconds. The production server is under the millisecond range and does not show a time except under heavy load.
1.a Get a test application server and production application server. Yes, you need both. The development server would be the developers workstation.
6.a Formalize the testing process by people other than the original developers.
6.b Write test cases.
6.c Do regression testing. Especially for "transperant to the user" changes.
8.a Have a naming standard for table columns. This will make your life SOOOO much easier.
11.a Where you do need hardcoded directories, externalize the locations in configuration files.
OS/2 did this.
OS/2 had (has?) in its file structure (HPFS) a concept known as extended attributes. I believe that the MAC also uses this (forked files).
Basically you can store anything you want in an extended attribute, then search for it within the file system utilities.
This is where meta-data and a database comes in. You need a large number of fields to describe what you are saving, then you can organize it (report) as you want with database queries.
Of course you need to actually enter all that information, but hey, I do NOT do content, that is the customer's problem...
Everyone is able to stop paying them whenever they want. You just have to do this funny thing called "follow the law."
Correct.
Except that most speed limits are set too low. It is assumed that most people will speed about 10/$units over the speed limit. Plus, they are set low for revenue generation.
With this system you will get a lot of slow drivers, increased time to get anywhere, and lots of frustrated drivers. And hopefully political pressure to raise limits.
I sometimes get to work at 140/km in a 100/km zone. Usually in a train of cars doing the same thing. The conditions make it safe: little traffic, lots of space between cars, dry roads, daylight.
24/7/365
If something is 24/7, how does adding the 365 make it more so? And what about leap years? Does 24/7/365 skip a day every four years?
<pet peave...>
I do research on virtual groups and social identity. Our names matter when we interact fora long time online. A name represents a person exclusively online.
Hmmmm, I set up a unique identity where I can. I must have over 20 different online personas.
So where I am RetroGeek here, I am something else on another board or newsgroup server.
Psych research in onthis question is showing the importance of our online names both to us as individuals and to the smooth running of social structures.
Ah, but I can be whoever I want to be. All I need to do is create another email address and register.
I find it amusing that there are so many Web services which insist upon you telling them a valid email address to complete some registration process. I set up an email address, give it to the service, they email me a registration code, I delete the email address. Mozilla (SeaMonkey?) saves the ui/pw for that site, so I do not need to remember them.
I am who I am, a different face to all......
While I wouldn't call maintenance coding simple, you seem to think it is much harder than it really is.
The real problem is getting all the interactions right. You tend to spend a lot of time finding out what code depends on what you are changing. Maybe that other code uses a workaround for buggy code you just fixed, which your fix will now break.
It is not the actual finding of bugs. It is tracing down all the interactions, and testing, testing, testing, testing.
The GP wanted to use in-house developers.
Agreeing with the bug-fixes though.....
The other problem is in creating a fork which only YOUR apps can use. So when the source owner (OSS or not) releases a fix, you have a problem on how to integrate it.....