I'm not sure how this is supposed to be "self-healing". If the exception isn't handled by somebody, you get a program crash. And this mechanism certainly isn't exclusive to Java; indeed any language which supports the notion of exceptions would certainly have a way to catch and deal with a thrown exception.
Perhaps Delerium can elaborate on what he means by "self-healing".
Java's stack traces for uncaught exceptions are handy though. I wish C++'s exception handling mechanism had something similiar built-in. I once read an article in Dr.Dobb's that showed you how to achieve the same thing in C++. Unfortunately it involved the use of ugly macros and required the programmer to remember to call these macros whenever you entered and left a function.
However, I bet you can write a C++ debugger that'll show you the exact location of where an uncaught exception was thrown thus alleviating the need for a stack trace. Anybody know of such a debugger?
> Yes, except in the case of derived classes -- the > writer of the derived class has to explicitly > write a dtor to ensure the parent class dtor is > called.
This is wrong. If your compiler requires this, then your compiler is broken.
The destructor in a base class has to be made virtual though to ensure that correct destructors will be called if you do something like this:
Base* o = new Derived(); delete o;
where Base is some class and Derived is derived from Base. If Base::~Base() is not virtual, then Derived::~Derived() would never get called.
I don't think the performance thing is the main isssue when using RTTI. The main issue when using RTTI is that whenever you write code that depends on the exact type of the object, you are no longer being object oriented. That is you are doing dynamic typing rather than dynamic binding. Dynamic typing == bad. Dynamic binding == good. This is probably what the parent poster learned in his OO class.
There are some instances where runtime type information is necessary (for example, you get some unknown object thats been deserialized from a network socket and you'll need to get it to the right type before you can do anything useful with it). But RTTI abuse is very easy to do, especially for the beginner.
> 1) WxWindows break OO principles by using message > maps; in other words, you don't overload a method, > you define which callback to call using an id > through a table. Message maps is a major reason > why MFC sucks.
You have two choices with respect to connecting events to event handlers in wxWidgets:
1. Use the aforementioned MFC-style message maps. 2. Use the Gtk/Qt signal/slot-style wxEvtHandler::Connect() method.
MFC-style message maps are more prevalent in the documentation, however. But if you don't wanna use them, you don't have to.
I', using the wxMozilla wxWindows component to be able to embed the Mozilla HTML rendering widget into an application I am working on. The downside is that it instantly adds a whopping 10MB to the whole package.
> In Debian, gaim has a dependency on NAS (Network > Audio System), so I'm forced to install it. I > don't need NAS. I don't want to install NAS. > Gentoo has a USE flag that allows me to declare > that I don't want anything to use NAS.
apt-get source gaim cd gaim-0.75 [edit debian/rules file, add "--disable-nas" to DEB_CONFIGURE_EXTRA_FLAGS which, incidently, is the default.. are you sure Debian's gaim depends on NAS?] debian/rules binary
viola... shiney new gaim deb which doesn't depend on NAS in gaim-0.75/..!
> relatively difficult to create a new.deb.
Not really.
I've never used Gentoo, but from what I understand it's claim to fame is that all packages are basically automagically installed from source. Bleh.. who has the time. Better to have most of everything installed from binary packages, and then selectively be able to *easily* rebuild a package from source. And in Debian, this is pretty easy to do.
> $150,000 is not enough for four years at my
> local comunity college.
Community colleges offer 4 year degrees?
At any rate, the community college I went to before I transferred to the university, irrc, offered courses at $50/credit. So a minimum full time workload (at least 12 credits) was about $600 for tuition. Add a couple of hundred for books (or maybe not.. I always waited until I was absolutely sure I needed the book), bringing the total up to maybe $800. Certainly not much more than $1000 per semester. I highly recommend to anybody whos strapped for cash to spend the first two years of college at the local junior college (just make sure all of your credits transfer to the 4 year school). You save a crapload of cash that way.
Plus, more than likely, all of your high school friends will be there too! So it'll seem like 13th and 14th grade.
> Dynamic typing is a Good Thing. It allows for > rapid development
Dynamic typing being a Good Thing depends on the context. Dynamic typing tends to move more bugs which could easily be caught at compile time to runtime. This means more testing needs to be done which actually drives up development costs and thus negates any benefit gained from "rapid development".
You could have the XML-RPC client make the XML-RPC call to the server, and simply not have the server return a response until it wants to send data back to the client. Of course, this will only work if the XML-RPC client and server implementations do not have a timeout value. Fortunately, most XML-RPC server implementations do not. You can use Apache's HTTP KeepAlive option to help implement this.
AFAIK, T-Mobile offers unlimited GPRS for $30/month if you don't have an existing voice plan or $20/month if you do. Just as soon as my AT&T Wireless contract is up, I'm switching.
I got one of those ultra-light Dell Latitude C400 notebooks. It's only about 3lbs and has only a 12" LCD screen, so it's very portable but yes it's ergonomics sucks (however, it's not *that* bad compared to other laptops I've had). Which is why I have the docking station for it so when I'm at my desk I can dock it and get the full fledged desktop PC experience (except for the noticeably slower hard drive compared to what you'd find in your average desktop).
And Dell's C/Dock II docking station has 2 PCI slots so if the USB doesn't give me enough expansion options, I can always go traditional and use the PCI slots.
I managed to get a good deal off Ebay and bought two C/Dock II's for $30 a piece so I have one for work and one for home. It's great being able to take my office machine to and from work without breaking my back in the process. At this point the 1.3Ghz CPU maybe somewhat on the slow side, but for my purposes it's great. The slow CPU speed means I can take longer coffee breaks in between compiles.
Not to go off on a tangent, but I wouldn't call MySQL a "proper relational database backend". It didn't even support foreign key constraints until InnoDB. But at any rate, I believe for a DBMS to be considered relational it must support the relational model which implies that it supports the concept of a transaction which (and correct if I'm wrong) MySQL does not.
I was able to unlock my Nokia 6800 which I got from AT&T by using one of those Nokia unlock code generator apps. At least, when I entered the code it said "Phone restriction off" on the phone. I haven't tried swapping in another SIM card yet tho to be 100% sure. Assuming this really worked, unlocking your Nokia is a breeze.
I dunno.. I really like my Nokia 6800. Flip out keyboard makes text messaging a breeze. It's also GPRS and Java enabled. Basic PDA functionality like an organizer and to-do list are included (or you can download and install other PDA-like J2ME apps). Only problem I've had with it so far is that the IrDA on it doesn't seem to like to play nice with the IrDA USB adapter I got for my desktop or the IrDA port on my laptop. And no bluetooth support. I finally ponied up for the data cable. $50 retail, but $30 off ebay. Still kind of expensive tho.
Re:cheapest mobile AIM + webbrowsing?
on
Death of the PDA?
·
· Score: 1
T-mobile is supposed to have unlimited GPRS for $30.00/month, if you don't have a voice plan, or $20.00/month if you do. With the $30/month deal, you can still do voice for $0.20 a minute tho.
I'm currently signed with AT&T with mLife/mMode. It sucks. I'm debating whether or not to just wait out the year (only signed with them for a year, thank $DEITY), or go ahead and have both T-Mobile and AT&T and just not use AT&T (my current mLife/mMode deal is $19/month for voice plus $2/month for mMode which gives me 0 bytes per month.. meaning that I pay $0.02/kb right away).
I've already racked up a $13 GPRS bill! And I've barely touched my minutes. It seems to me that wireless internet is far more useful to me than voice. I wish I had done more research:-(.
Slashdot Mirror
Or he could try using XPLite to take out the crap he doesn't need.
> "Evidence" can often be made to support any number
> of theories
Thats why we have peer review.
I'm not sure how this is supposed to be "self-healing". If the exception isn't handled by somebody, you get a program crash. And this mechanism certainly isn't exclusive to Java; indeed any language which supports the notion of exceptions would certainly have a way to catch and deal with a thrown exception.
Perhaps Delerium can elaborate on what he means by "self-healing".
Java's stack traces for uncaught exceptions are handy though. I wish C++'s exception handling mechanism had something similiar built-in. I once read an article in Dr.Dobb's that showed you how to achieve the same thing in C++. Unfortunately it involved the use of ugly macros and required the programmer to remember to call these macros whenever you entered and left a function.
However, I bet you can write a C++ debugger that'll show you the exact location of where an uncaught exception was thrown thus alleviating the need for a stack trace. Anybody know of such a debugger?
> Yes, except in the case of derived classes -- the
> writer of the derived class has to explicitly
> write a dtor to ensure the parent class dtor is
> called.
This is wrong. If your compiler requires this, then your compiler is broken.
The destructor in a base class has to be made virtual though to ensure that correct destructors will be called if you do something like this:
Base* o = new Derived();
delete o;
where Base is some class and Derived is derived from Base. If Base::~Base() is not virtual, then Derived::~Derived() would never get called.
I don't think the performance thing is the main isssue when using RTTI. The main issue when using RTTI is that whenever you write code that depends on the exact type of the object, you are no longer being object oriented. That is you are doing dynamic typing rather than dynamic binding. Dynamic typing == bad. Dynamic binding == good. This is probably what the parent poster learned in his OO class.
There are some instances where runtime type information is necessary (for example, you get some unknown object thats been deserialized from a network socket and you'll need to get it to the right type before you can do anything useful with it). But RTTI abuse is very easy to do, especially for the beginner.
I like cl.exe a little bit better mainly because it compiles a hell of a lot faster than gcc. At least time I tried using the MinGW suite of tools.
> 1) WxWindows break OO principles by using message
> maps; in other words, you don't overload a method,
> you define which callback to call using an id
> through a table. Message maps is a major reason
> why MFC sucks.
You have two choices with respect to connecting events to event handlers in wxWidgets:
1. Use the aforementioned MFC-style message maps.
2. Use the Gtk/Qt signal/slot-style wxEvtHandler::Connect() method.
MFC-style message maps are more prevalent in the documentation, however. But if you don't wanna use them, you don't have to.
> can't be destroyed in the way MS did to
> Netscape
I think Netscape didn't need Microsoft's help. They destroyed themselves pretty well on their own.
I', using the wxMozilla wxWindows component to be able to embed the Mozilla HTML rendering widget into an application I am working on. The downside is that it instantly adds a whopping 10MB to the whole package.
> In Debian, gaim has a dependency on NAS (Network
.. are you sure Debian's gaim depends on NAS?]
... shiney new gaim deb which doesn't depend on NAS in gaim-0.75/..!
.deb.
.. who has the time. Better to have most of everything installed from binary packages, and then selectively be able to *easily* rebuild a package from source. And in Debian, this is pretty easy to do.
> Audio System), so I'm forced to install it. I
> don't need NAS. I don't want to install NAS.
> Gentoo has a USE flag that allows me to declare
> that I don't want anything to use NAS.
apt-get source gaim
cd gaim-0.75
[edit debian/rules file, add "--disable-nas" to DEB_CONFIGURE_EXTRA_FLAGS which, incidently, is the default
debian/rules binary
viola
> relatively difficult to create a new
Not really.
I've never used Gentoo, but from what I understand it's claim to fame is that all packages are basically automagically installed from source. Bleh
> $150,000 is not enough for four years at my
.. I always waited until I was absolutely sure I needed the book), bringing the total up to maybe $800. Certainly not much more than $1000 per semester. I highly recommend to anybody whos strapped for cash to spend the first two years of college at the local junior college (just make sure all of your credits transfer to the 4 year school). You save a crapload of cash that way.
> local comunity college.
Community colleges offer 4 year degrees?
At any rate, the community college I went to before I transferred to the university, irrc, offered courses at $50/credit. So a minimum full time workload (at least 12 credits) was about $600 for tuition. Add a couple of hundred for books (or maybe not
Plus, more than likely, all of your high school friends will be there too! So it'll seem like 13th and 14th grade.
> Dynamic typing is a Good Thing. It allows for
> rapid development
Dynamic typing being a Good Thing depends on the context. Dynamic typing tends to move more bugs which could easily be caught at compile time to runtime. This means more testing needs to be done which actually drives up development costs and thus negates any benefit gained from "rapid development".
Look up Flyweight design pattern.
As if this annoying feature was strictly limited to Windows server software...
Now if you had said "commercial server software", that would've much more sense.
You could have the XML-RPC client make the XML-RPC call to the server, and simply not have the server return a response until it wants to send data back to the client. Of course, this will only work if the XML-RPC client and server implementations do not have a timeout value. Fortunately, most XML-RPC server implementations do not. You can use Apache's HTTP KeepAlive option to help implement this.
AFAIK, T-Mobile offers unlimited GPRS for $30/month if you don't have an existing voice plan or $20/month if you do. Just as soon as my AT&T Wireless contract is up, I'm switching.
Except the CoCo's ROM BASIC was made by Microsoft.
OOPS!
Does this thing do Swing only? The Shockwave demos seems to indicate "yes".
Just get a docking station.
I got one of those ultra-light Dell Latitude C400 notebooks. It's only about 3lbs and has only a 12" LCD screen, so it's very portable but yes it's ergonomics sucks (however, it's not *that* bad compared to other laptops I've had). Which is why I have the docking station for it so when I'm at my desk I can dock it and get the full fledged desktop PC experience (except for the noticeably slower hard drive compared to what you'd find in your average desktop).
And Dell's C/Dock II docking station has 2 PCI slots so if the USB doesn't give me enough expansion options, I can always go traditional and use the PCI slots.
I managed to get a good deal off Ebay and bought two C/Dock II's for $30 a piece so I have one for work and one for home. It's great being able to take my office machine to and from work without breaking my back in the process. At this point the 1.3Ghz CPU maybe somewhat on the slow side, but for my purposes it's great. The slow CPU speed means I can take longer coffee breaks in between compiles.
> proper relational database backend like MySQL
Not to go off on a tangent, but I wouldn't call MySQL a "proper relational database backend". It didn't even support foreign key constraints until InnoDB. But at any rate, I believe for a DBMS to be considered relational it must support the relational model which implies that it supports the concept of a transaction which (and correct if I'm wrong) MySQL does not.
I was able to unlock my Nokia 6800 which I got from AT&T by using one of those Nokia unlock code generator apps. At least, when I entered the code it said "Phone restriction off" on the phone. I haven't tried swapping in another SIM card yet tho to be 100% sure. Assuming this really worked, unlocking your Nokia is a breeze.
I dunno .. I really like my Nokia 6800. Flip out keyboard makes text messaging a breeze. It's also GPRS and Java enabled. Basic PDA functionality like an organizer and to-do list are included (or you can download and install other PDA-like J2ME apps). Only problem I've had with it so far is that the IrDA on it doesn't seem to like to play nice with the IrDA USB adapter I got for my desktop or the IrDA port on my laptop. And no bluetooth support. I finally ponied up for the data cable. $50 retail, but $30 off ebay. Still kind of expensive tho.
The Panasonic Toughbook 18 is one.
T-mobile is supposed to have unlimited GPRS for $30.00/month, if you don't have a voice plan, or $20.00/month if you do. With the $30/month deal, you can still do voice for $0.20 a minute tho.
.. meaning that I pay $0.02/kb right away).
:-(.
I'm currently signed with AT&T with mLife/mMode. It sucks. I'm debating whether or not to just wait out the year (only signed with them for a year, thank $DEITY), or go ahead and have both T-Mobile and AT&T and just not use AT&T (my current mLife/mMode deal is $19/month for voice plus $2/month for mMode which gives me 0 bytes per month
I've already racked up a $13 GPRS bill! And I've barely touched my minutes. It seems to me that wireless internet is far more useful to me than voice. I wish I had done more research
> It'll probably be years before I'll be able to
> run my own software on my own phone.
My Nokia 6800 lets me upload J2ME apps of my own design into it.