- Feature Articles
- CodeSOD
- Error'd
- Forums
-
Other Articles
- Random Article
- Other Series
- Alex's Soapbox
- Announcements
- Best of…
- Best of Email
- Best of the Sidebar
- Bring Your Own Code
- Coded Smorgasbord
- Mandatory Fun Day
- Off Topic
- Representative Line
- News Roundup
- Editor's Soapbox
- Software on the Rocks
- Souvenir Potpourri
- Sponsor Post
- Tales from the Interview
- The Daily WTF: Live
- Virtudyne
Admin
Admin
And what in this story suggests than any of the bolded parts above apply to this situation? It's a one-off app that will be used for a couple of weeks, isn't critical, and will only be used by a handful of users at a time.
Admin
Holy crap. T0pCod3r, or whoever's running that account these days, TAKE NOTE: Bim Job just schooled the fuck outta TDWTF trolls.
Nailed it, mothafucka: Now sing your win!Admin
Admin
Admin
Addendum (2010-03-18 01:44): It may turn out to be blue cheese (yummy!) in the end, but it still initially stinks.
Admin
There's no way to interactively do anything on a spreadsheet that's thousands of columns wide. You lose key benefits of the spreadsheet: you have to type addresses manually (good luck selecting exactly 365 columns with a mouse), you need code written to ensure that the cells adhere to your stencil/template, you can't get meaningful diffs, and so on. By the time you're done making it robust, you end up with so much non-visual scripting that you might have as well done everything in some scripting language to begin with.
Spreadsheets are visual, purely-functional programming paradigms. They are the living proof that people can think in terms of purely functional programs. That's the key power of a spreadsheet: a visual, purely functional environment. If you are writing a lot of scripts around them, the benefits evaporate very, very quickly. A spreadsheet with controls in it is the first sign that the project needs to be transferred to a better environment.
Myself, I'd probably do it all in Octave, if you're into open source solutions. There are way more powerful optimizers for Octave/Matlab than Solver, it's easy to get data from pretty much any source, and so on.
Alternatively, something like Ocaml might be useful if you're into keeping it robust and not prone to errors.
I know that many people who do finances and trading use Excel a lot, but that doesn't mean it's a good thing. I know a guy who was dealing with the fallout from those spreadsheets in Worldcom. Try getting some Joe Schmoe to notice that a few formulas somewhere on sheet #5 were copied/pasted wrongly to sheet #3, and they don't do their job anymore. Good luck.
Admin
First one is getting the thing to compile -- it got better over the years, but still it makes compiling gcc look not all that bad. Then wrapping your mind around enough things to make even the slightest change. Something like changing the order of two buttons in a dialog box. Then adding a new button or another control and having it react to events...
I'd say that lack of real documentation for the self-rolled framework they use is the nail in the coffin. Nowhere does it answer "why" something was done this-a-way and not that-a-way. If you think about implementing a new feature, and there's nothing "close enough" -- you will have no mental image of how to use their framework to do what you need. It's easy enough to understand the individual APIs, but unless you know where to look first, you may never find them.
There's no good reason why an open source project on the scale of OO shouldn't be documented enough to have any decent programmer be able to pick up the basics in a couple of days. The difference is that in case of say .NET, Qt or even Java frameworks, there's plenty of documentation that tells you why thing are the way they are, and where to look for certain functionality. For OO, nothing like that exists IIRC. Good luck hacking on it if you don't wish to spend weeks on doing so.
I have a bunch of relatively simple things I'd like to change, and I know it'd take maybe 20-30 minutes to change them in anything that uses Qt/KDE or WPF+.NET. In OO, it'd take me that long merely to figure that there's no real developer documentation. Their DevGuide may as well be written it Chinese. Surely it'd make perfect sense to those who know Chinese. The way it stands, the "First Steps" chapter begins thusly:
Helpful as hell. So many words, so little content. The DevGuide is supposedly useful if you want to write add-ons etc. To hack on OO itself it's virtually useless at first, and there's no other real document that would do the job. Some Developer's Guide, my ass.Contrast that with thoughtful tutorials and strategic intros provided with Qt or .NET. OO's developer wiki looks like something that a bunch of monkeys threw together...
I'm a happy OO user, but the fact that it's open source is absolutely meaningless to me. The barrier to entry is needlessly big. To me, OO is freeware, and unless some hard working soul comes up with an "OO hacking" book for real people, with some real and useful examples (not all toy ones, like reversing a menu item), it won't change.
Reading this reinforces the barrier to entry: not only there's no real documentation, but the architecture itself is so counterproductive that I don't really blame anyone who refuses to come anywhere near that codebase. Obviously having just one string class in the project is never enough. I just imagine what's waiting if you dig any deeper than that.
It's a big feat that it works as well as it does, and I'm obviously a happy camper, using both OO.org and NeoOffice. But that's it. I'd pick away on gcc codebase any day over picking away on OO. And I guess that says something IMHO how comparably clean gcc is, even though -- again, purely IMHO -- on a barrier-to-entry scale, gcc scores pretty big.
Admin
Citations for articles written online please.
Admin
Admin
There's a little thing called a 'race condition' you might want to look that up.
If there is ANY chance that two people could hit the submit button close enough together to cause a problem Murphy says it'll happen and you will get corrupted data.
Admin
Well you see in a properly functioning free market the price will be pushed towards the marginal cost of a good (ie the $0.01 for pressing the CD), the development costs do not matter. This is why monopolies are bad for consumers, they keep prices high, stifle innovation (how long was IE6 the current version, with effectively no change, more than a decade?))
Admin
'Beardy' probably still applies...
PHP yes, it was hard to write secure code in it... (I never did, but I read about the things it used to do brrr) you could always just write it in perl or something...
MySQL... well it should be hidden from evil people anyway and it'd be better than Access :P (and if you don't need REALLY high concurrency it'd be just fine)
Apache was probably pretty solid.
Admin
I dunno about that, depends on how much setup you have to do, if you have some infrastructure in place already you could get it to the point where you just create the tables and upload your php... (like it is on most commercial webhosts)
Admin
I doubt it's impossible, you could almost certainly write a macro in OOBasic to do it
Admin
Hey now MySQL is a real grown up RDBMS now, and has been for some time, didn't really used to be, but times change.
Admin
FTFY
Admin
Admin
OMG I am so sick an tired of this argument that OSS is the best thing ever. Use what the hell you want to get the job done.... are you lot that hard up for a life you have to trawl through sites like this just to add another witty anecdote about "M$ is teh suxor man!"
Grow up and start focusing on getting the job done.
Admin
[email protected]