SQL Server has a power, flexibility and ease-of-use which enable it to be used as a platform for applications for which any other database server would be completely inappropriate. It's like MS Office in this respect: you can build intricate systems in Excel VBA (and many business users do...), and it's only when those systems have to scale up, show true resilience, or adapt to new requirements that the inadequacy of Office as a platform for applications development is revealed.
With SQL Server, you can push a whole lot of business logic onto the server via T-SQL and stored procedures, and leave the VB monkeys to write cute GUI front-ends for the resulting behemoth. This is really seductive - as a developer, I love the way T-SQL lets you build what is in effect an API to a database, with named procedures and parameters offering a gateway to an encapsulated data model (the actual tables, relationships and queries that service the client's requests). But beyond a certain point this whole model of development (client-server, with business logic wrapped up in triggers and stored procedures and a monkeyed-up GUI servicing the end-users) is dangerously limiting - you're too closely bound to the platform, and the platform can only carry so much weight.
SQL Server is clever in ways that are often very helpful, but the minute it stops you thinking of the persistence layer of your applications as essentially dumb, and dumb for a good reason - the same reason you don't keep your brains in your stomach - you start saving up troubles for the future. At least with MySQL you'd never be tempted to make that mistake...
Overall I'm a fan of SQL Server, which I agree is one of the best things MS have done; but it needs judicious handling, else you'll end up married to the beast.
I have a humanities degree, and no previous experience of working in a techie job (unless you count a few weeks' coding I did for a small firm when I was 16). There is no way I would ever have got my CV looked at by any potential tech employer, no matter how much I put on my CV about how I was programming at age 10, run linux at home, yadda-yadda-yadda...
I got my current job, which involves writing end-user apps and back-end processes for a small UK bank, because I started doing those things in VBA while I was working in admin, and impressed a couple of the right people. That's all. Nobody interviewed me for my current role. Nobody looked at my CV. I was about to get the boot as I'd automated my own job out of existence, and at just about the last minute someone with a new department to staff invented a job and gave it to me.
Who knows you're good? Who knows someone who's looking for someone they've heard is good, and doesn't want to try to sift through hundreds of resumes full of buzzwords and bullshit to find it? Recently my department employed a new graduate. His paper skills were actually pretty cool, but the fact that he was a friend of one of our existing developers surely didn't hurt...
Test-first is the bit of XP I find it most useful to employ in a non-XP environment (the one I work in). I don't know if it's unique to XP, but kudos to Kent & co for popularising the approach. Having said that, I've spent the past two days knocking something together without the aid of unit tests, just because the something in question was a VB IDE add-in and I could really see no way to test it as it stood using VBUnit. Sometimes you run up against the limitations of your tools (with VB, that's quite a lot of the time in fact...)
In what sense is Gator being blocked, and by what means? If the means employed are threats of legal action, then there's a case for the EFF to fight and there are strong reasons for them to fight it.
Gator changes what happens when you visit certain websites. So, in some cases, does Mozilla with pop-ups disabled. If I want to download a bunch of HTML source from a URL using my own custom application, process it through a Swedish Chef dialectizer and display it on my desktop in 48pt Courier over an animated backdrop of Mr Hankey the Christmas Poo, I can (and maybe I will...later...).
The end-user de facto controls the appearance of HTML-formatted content. The idea that content-owners should be able to use legal means to force me to view their pop-ups, or refrain from opening my own over their content, seems and is downright wrong. It's like being told not to look in a shop-window whilst wearing spectacles with pictures of naked choirboys taped to the insides, and you know how much I hate it when that happens.
If, on the other hand, Gator is being blocked by technical means, then fair enough. There's no reason why content-owners should have to create content that can be rendered in Hankey-o-Bork-o-Vision, and if they can detect and refuse to respond to requests coming from my custom client (or some Gator-"enabled" browser), then they're surely entitled to do so. That's true even if it's Mozilla they're blocking. More fool them if they do, but they're within their rights: they can configure their server and write their CGI any damn way they want.
Re:Doing OOP? You want this book.
on
Design Patterns
·
· Score: 1
Design Patterns may be one of those books that is useful to people in different ways at different stages in their career.
I read it a couple of months ago, during a holiday in Ireland with my family (who mocked me righteously for doing so). Absorbing the book away from a computer screen is not a bad idea, since one of the ways it can be useful is as an aid to consolidating the things you've been learning about programming during the six months you just spent chained to a keyboard.
I expect to return to it again soon, this time as a reference. A lot of the GoF patterns don't really fly in Visual Basic, which is what I use at work - Singleton for instance just degenerates into Global Variable, its evil twin. But I'm certainly seeing the merits of object composition, not least because implementation inheritance has to be done through delegation in VB6 anyway.
The extent to which a more experienced coder will find such a tome useful probably depends on that individual's propensity for holistic thinking. If you spend all your time rooting around among the trees, the aerial view of the wood may surprise you. Others may have been in the habit of drawing patterns out of their own experience for years; even so, it can be good to see them named and catalogued, and to have a common terminology for communicating with others about them.
The older KDE seemed a lot less demanding - I also ran it without trouble, a couple of years ago. I don't know when KDE started using artsd. Memory usage might be another issue: 64Mb of RAM doesn't seem like quite so much nowadays.
Am I right in thinking that using artsd to manage its multimedia is one of the big reasons why KDE3 is so fricken' slow on my old PIII 450Mhz? Or is that just KDE3 in general?
If it's not really artsd's fault, then I might start using it again, although I don't think I'll be going back to KDE from Gnome until I have a much faster computer (and possibly not even then).
Veering even further off-topic: will artsd let you record and playback at the same time (with, say, Audacity)? This is something I miss, really badly, since I switched from Windows...
I've tried and failed to use Anjuta. Out of the box, on my RedHat 7.3 system, it simply refused to play ball, insisting that it couldn't find things like glade or the gtk libraries that I know for a fact are there. It's fair enough if the app needed me to tell it where those things were, but the sticking point is that it didn't even ask: it behaved as if everything should happen automagically, and offered no way of fixing things when they didn't. Maybe a couple of releases down the line I'll go back to it and try again...
The Wrox P2P title Beginning Linux Programming, by Richard Stones and Neil Matthew, covers some similar ground: makefiles, shell scripts, some basic C, Perl and TCL (not really basic - they show you how to use these languages to do some typical linux-y things, rather than explain about structs and scalars), an introduction to GTK and a chapter at the end on writing device drivers. If you're an intermediate-level programmer unfamiliar with Linux as a programming environment, this book offers a good way in; recommended for VB weenies with a conscience.
You can write ASP in JavaScript (well, OK, JScript) if you wanna. And you can use it as a scripting language for COM components of any kind, if you use the MS Script Control. I've scripted terminal emulation clients and MS Office apps in JScript. I don't know if there's an equivalent framework in *nix - is th'Gimp ECMA-scriptable? How about Gnumeric?
In MS land, you're probably better off with Ruby or Python plus COM bindings, if your place of work will let you use them (mine won't...). But ECMA-script is a pretty neat second best - unless you'd rather use VBScript...
I don't think HTML is hard or complicated at all. Good website design is hard, and giving good usability is very hard indeed. If you're worrying about the right things, then the actual mark-up should be the least of your worries.
Really the two biggest problems I've had with VB have been the following:
Inability to pass functions as parameters. Although VB6 has an event-handling mechanism that takes some of the pain out of not being able to write little anonymous functions as callback handlers, there's still a dull ache right there where that feature should be. Plus, right now I'm playing with Ruby and every extra minute I spend with that language's facility with iterators and blocks makes me hate VB more.
No threads. You can spawn ActiveX servers, get them to run off a timer pulse on a form, and pass them work to do "in the background", but it's a rediculously cumbersome way of solving a very common programming problem.
I don't find VB's strength to be in GUI-building especially, although it does make that almost insultingly easy. For my purposes, it's best as a COM client providing glue and business logic for other components which are doing most of the real "execution". But as a language - bleh! If my place of work could be a little more open-minded about alternatives such as Python, Ruby (both of which can also work as COM clients, and even servers), I'd be a much happier bunny.
VCC isn't an option for me simply because I find the idea of having a Wizard dump loads of boilerplate code all over an application simply offensive. I thought the point of inheritance was to remove the need for cut-and-paste duplication of this kind - isn't it supposed to clear all that crap out of the way?
I had a similar experience rendering bits of the Mandelbrot set at low resolution, 128 iterations max, using BBC Basic on a BBC micro. That took all night. Then did the same on an Acorn Archimedes, using ARM assembler. That happened in front of your eyes. The same images can now be rendered pretty much instantaneously...
Victims? Don't be
melodramatic. Look down there. Would you feel any pity if one of those dots stopped moving
forever? If I offered you 20,000 pounds for every dot that stopped moving, would you really, old
man, tell me to keep my money? Or would you calculate how many dots you could afford to
spare? Free of income tax, old man. Free of income tax. It's the only way to save money
nowadays...
Slashdot Mirror
SQL Server has a power, flexibility and ease-of-use which enable it to be used as a platform for applications for which any other database server would be completely inappropriate. It's like MS Office in this respect: you can build intricate systems in Excel VBA (and many business users do...), and it's only when those systems have to scale up, show true resilience, or adapt to new requirements that the inadequacy of Office as a platform for applications development is revealed.
With SQL Server, you can push a whole lot of business logic onto the server via T-SQL and stored procedures, and leave the VB monkeys to write cute GUI front-ends for the resulting behemoth. This is really seductive - as a developer, I love the way T-SQL lets you build what is in effect an API to a database, with named procedures and parameters offering a gateway to an encapsulated data model (the actual tables, relationships and queries that service the client's requests). But beyond a certain point this whole model of development (client-server, with business logic wrapped up in triggers and stored procedures and a monkeyed-up GUI servicing the end-users) is dangerously limiting - you're too closely bound to the platform, and the platform can only carry so much weight.
SQL Server is clever in ways that are often very helpful, but the minute it stops you thinking of the persistence layer of your applications as essentially dumb, and dumb for a good reason - the same reason you don't keep your brains in your stomach - you start saving up troubles for the future. At least with MySQL you'd never be tempted to make that mistake...
Overall I'm a fan of SQL Server, which I agree is one of the best things MS have done; but it needs judicious handling, else you'll end up married to the beast.
I second this.
I have a humanities degree, and no previous experience of working in a techie job (unless you count a few weeks' coding I did for a small firm when I was 16). There is no way I would ever have got my CV looked at by any potential tech employer, no matter how much I put on my CV about how I was programming at age 10, run linux at home, yadda-yadda-yadda...
I got my current job, which involves writing end-user apps and back-end processes for a small UK bank, because I started doing those things in VBA while I was working in admin, and impressed a couple of the right people. That's all. Nobody interviewed me for my current role. Nobody looked at my CV. I was about to get the boot as I'd automated my own job out of existence, and at just about the last minute someone with a new department to staff invented a job and gave it to me.
Who knows you're good? Who knows someone who's looking for someone they've heard is good, and doesn't want to try to sift through hundreds of resumes full of buzzwords and bullshit to find it? Recently my department employed a new graduate. His paper skills were actually pretty cool, but the fact that he was a friend of one of our existing developers surely didn't hurt...
Test-first is the bit of XP I find it most useful to employ in a non-XP environment (the one I work in). I don't know if it's unique to XP, but kudos to Kent & co for popularising the approach. Having said that, I've spent the past two days knocking something together without the aid of unit tests, just because the something in question was a VB IDE add-in and I could really see no way to test it as it stood using VBUnit. Sometimes you run up against the limitations of your tools (with VB, that's quite a lot of the time in fact...)
I don't understand this story.
In what sense is Gator being blocked, and by what means? If the means employed are threats of legal action, then there's a case for the EFF to fight and there are strong reasons for them to fight it.
Gator changes what happens when you visit certain websites. So, in some cases, does Mozilla with pop-ups disabled. If I want to download a bunch of HTML source from a URL using my own custom application, process it through a Swedish Chef dialectizer and display it on my desktop in 48pt Courier over an animated backdrop of Mr Hankey the Christmas Poo, I can (and maybe I will...later...).
The end-user de facto controls the appearance of HTML-formatted content. The idea that content-owners should be able to use legal means to force me to view their pop-ups, or refrain from opening my own over their content, seems and is downright wrong. It's like being told not to look in a shop-window whilst wearing spectacles with pictures of naked choirboys taped to the insides, and you know how much I hate it when that happens.
If, on the other hand, Gator is being blocked by technical means, then fair enough. There's no reason why content-owners should have to create content that can be rendered in Hankey-o-Bork-o-Vision, and if they can detect and refuse to respond to requests coming from my custom client (or some Gator-"enabled" browser), then they're surely entitled to do so. That's true even if it's Mozilla they're blocking. More fool them if they do, but they're within their rights: they can configure their server and write their CGI any damn way they want.
Design Patterns may be one of those books that is useful to people in different ways at different stages in their career.
I read it a couple of months ago, during a holiday in Ireland with my family (who mocked me righteously for doing so). Absorbing the book away from a computer screen is not a bad idea, since one of the ways it can be useful is as an aid to consolidating the things you've been learning about programming during the six months you just spent chained to a keyboard.
I expect to return to it again soon, this time as a reference. A lot of the GoF patterns don't really fly in Visual Basic, which is what I use at work - Singleton for instance just degenerates into Global Variable, its evil twin. But I'm certainly seeing the merits of object composition, not least because implementation inheritance has to be done through delegation in VB6 anyway.
The extent to which a more experienced coder will find such a tome useful probably depends on that individual's propensity for holistic thinking. If you spend all your time rooting around among the trees, the aerial view of the wood may surprise you. Others may have been in the habit of drawing patterns out of their own experience for years; even so, it can be good to see them named and catalogued, and to have a common terminology for communicating with others about them.
The older KDE seemed a lot less demanding - I also ran it without trouble, a couple of years ago. I don't know when KDE started using artsd. Memory usage might be another issue: 64Mb of RAM doesn't seem like quite so much nowadays.
Am I right in thinking that using artsd to manage its multimedia is one of the big reasons why KDE3 is so fricken' slow on my old PIII 450Mhz? Or is that just KDE3 in general?
If it's not really artsd's fault, then I might start using it again, although I don't think I'll be going back to KDE from Gnome until I have a much faster computer (and possibly not even then).
Veering even further off-topic: will artsd let you record and playback at the same time (with, say, Audacity)? This is something I miss, really badly, since I switched from Windows...
I've tried and failed to use Anjuta. Out of the box, on my RedHat 7.3 system, it simply refused to play ball, insisting that it couldn't find things like glade or the gtk libraries that I know for a fact are there. It's fair enough if the app needed me to tell it where those things were, but the sticking point is that it didn't even ask: it behaved as if everything should happen automagically, and offered no way of fixing things when they didn't. Maybe a couple of releases down the line I'll go back to it and try again...
The Wrox P2P title Beginning Linux Programming, by Richard Stones and Neil Matthew, covers some similar ground: makefiles, shell scripts, some basic C, Perl and TCL (not really basic - they show you how to use these languages to do some typical linux-y things, rather than explain about structs and scalars), an introduction to GTK and a chapter at the end on writing device drivers. If you're an intermediate-level programmer unfamiliar with Linux as a programming environment, this book offers a good way in; recommended for VB weenies with a conscience.
You can write ASP in JavaScript (well, OK, JScript) if you wanna. And you can use it as a scripting language for COM components of any kind, if you use the MS Script Control. I've scripted terminal emulation clients and MS Office apps in JScript. I don't know if there's an equivalent framework in *nix - is th'Gimp ECMA-scriptable? How about Gnumeric?
In MS land, you're probably better off with Ruby or Python plus COM bindings, if your place of work will let you use them (mine won't...). But ECMA-script is a pretty neat second best - unless you'd rather use VBScript...
I don't think HTML is hard or complicated at all. Good website design is hard, and giving good usability is very hard indeed. If you're worrying about the right things, then the actual mark-up should be the least of your worries.
Really the two biggest problems I've had with VB have been the following:
I don't find VB's strength to be in GUI-building especially, although it does make that almost insultingly easy. For my purposes, it's best as a COM client providing glue and business logic for other components which are doing most of the real "execution". But as a language - bleh! If my place of work could be a little more open-minded about alternatives such as Python, Ruby (both of which can also work as COM clients, and even servers), I'd be a much happier bunny.
VCC isn't an option for me simply because I find the idea of having a Wizard dump loads of boilerplate code all over an application simply offensive. I thought the point of inheritance was to remove the need for cut-and-paste duplication of this kind - isn't it supposed to clear all that crap out of the way?
What's in the box? Pain...
I had a similar experience rendering bits of the Mandelbrot set at low resolution, 128 iterations max, using BBC Basic on a BBC micro. That took all night. Then did the same on an Acorn Archimedes, using ARM assembler. That happened in front of your eyes. The same images can now be rendered pretty much instantaneously...
First person to source this quote gets a no-prize
Victims? Don't be melodramatic. Look down there. Would you feel any pity if one of those dots stopped moving forever? If I offered you 20,000 pounds for every dot that stopped moving, would you really, old man, tell me to keep my money? Or would you calculate how many dots you could afford to spare? Free of income tax, old man. Free of income tax. It's the only way to save money nowadays...