Category Archives: Code

MySchool: so wrong

Background: The Australian federal government has finally pushed out a web site publishing performance metrics for all schools throughout Australia. There has been much brouhaha regarding this. For some reason, the go-live wasn’t a quiet one, but a very loud, flick-on-the-switch big-bang go live.

Naturally, the website asploded.

Any website that’s going to be hit by 1% of the Australian population the moment it goes live is going to blow up unless there are some cluey, experienced people behind it. Clever, inexperienced people, or experienced idiots with a large budget might stand a chance if things got progressively worse over time, but turn it on and hammer it on day one? MySchool.edu.au does not have cluey, experienced people behind it. There are various signs.

For a start, what is it with the TLD? .edu.au seems fine, but what’s wrong with a redirect from .gov.au given they were the folks running around promoting it? It’s not like myschool is an education institution.

Then you get there. Guess what? It won’t work without JavaScript. At all. Because typing in a string and hitting enter demands the availability of JavaScript. Using <form> is so 2000s. Get with the new decade! It’s so vital to the site that users must not be allowed in if they don’t have JavaScript. Screw the blind! They’ve only got one school to go to anyway.

And the site is slow, amazing slow. But I guess if you’ve got to download all that JavaScript to enter that string, of course it’s going to be slow. Switching to a different set of data? Couldn’t download that and just do a hide/show, no you’ve got to do some kinda AJAX-y postback crap for a massive round-trip delay; if you were dealing with rapidly changing data, that might almost make sense; every year this website will get data updates, so no: this makes no sense. I clicked on it, and a long time later, something happened to the web page. In the meantime, I went off to get a drink. Alternatively, you could just show a table for each year, and skip the damn JavaScript altogether. Why there’s even a backend is beyond me, this whole thing could be served perfectly well – and mind-numbingly quickly – from static pages.

And for the purpose it’s intended for: parents picking a school for their kids. Can you compare schools? No. Open them up in different browser tabs, if you have a tabbed browser (remember: the blind can go take a flying leap). Good thing the site is chocked full of JavaScript. And the JavaScript is used for handy things like map-based locating of schools, and – oh, hang on, no it’s not. There’s no Google-maps mash-up. Good thing the site is chocked full of JavaScript.

Clearly, the entire site has been an exercise in some programmer somewhere bolstering their resume rather than giving the client something appropriate. Either that, or a manager was in charge of the feature spec, and demanded all the latest buzzwords that they had heard but didn’t understand. I’m betting it took more than a year to build. Feel free to speculate.

I’m also willing to bet the price on this site was more than the $50,000 it should have cost (one person, three months). I’m imagining about two or three orders of magnitude more. I’m figuring the servers required for this aren’t running in some guy’s bedroom, even though that would be about all that’s required for such a simple dataset that’s presented in such a straightforward way.

Must try harder.

One third party: a tiny bit more complexity; Many: :-(

I didn’t find the most interesting part of the StackOverflow podcast #16 to be my question. I direct you to 17:06 in, where Jeff discusses the pros and cons of using OpenID as the authentication mechanism for StackOverflow:

Atwood: Granted, there’s a third entity here so there’s going to be a tiny bit more complexity.

What Jeff’s overlooked here is the Combinatorial Complexity; he’s not hooking up with an OpenID provider, he’s hooking up with all OpenID providers, which he acknowledged earlier can be a bit of a problem [34:48 in podcast #7]:

Atwood: Well you can, I found that Yahoo doesn’t really do attribute exchange very well.

If you look at the uservoice… forum… bug-reporting… suggestion-y thing for the StackOverflow beta, you see a lot of people complaining “my OpenID provider doesn’t work [at all]/[properly] with your site”.
OpenID has a spec, but given the difficulties being experienced, it mustn’t be terribly tight or there’s no reference implementation to validate against.

Having said all that, guess what I’m going to be using as my authentication process on my next website?

Linker problems galore

LINK : fatal error LNK1104: cannot open file ‘mfcs42d.lib’

Go to Tools | Options… | Projects and Solutions | VC++ Directories, select Show Directories for: Library files and insert the path to your existing mfcs42d.lib.

Now you get
LNK2019: unresolved external symbol “public: __thiscall AFX_MODULE_STATE::AFX_MODULE_STATE(int,long (__stdcall*)(struct HWND__ *,unsigned int,unsigned int,long),unsigned long)” (??0AFX_MODULE_STATE@@QAE@HP6GJPAUHWND__@@IIJ@ZK@Z) referenced in function “public: __thiscall _AFX_DLL_MODULE_STATE::_AFX_DLL_MODULE_STATE(void)” (??0_AFX_DLL_MODULE_STATE@@QAE@XZ)

Remove mfcs42d.lib and you get
mfcs90d.lib(dllmodul.obj) : error LNK2005: _DllMain@12 already defined in MSVCRTD.lib(dllmain.obj)

Know what you did wrong?

I’ve got the answer.

Don’t link to mfcs42d (and, for that matter, the non-debug version mfcs42.lib) in your 2008 project. If that name’s hard-coded into your project – and it is, isn’t it? – then change the name to mfcs90d and you’ll be fine.

Visual C++ Compiler Error C2316: Make with the class definition

Visual C++ Compiler Error C2316

You know, I was getting this error when porting some exception catching code from Visual C++ 6 to version 9. The error message helpfully states that 'exception' cannot be caught as the destructor and/or copy constructor are inaccessible

Weird thing was, it was wrong. Both the copy constructor and destructor for the class and all it’s ancestors were in public scope.

Eventually there was the figuring out: the exception was being anonymously caught, like so:
try {
whatever();
}
catch (MyExceptionClass&) {}
And to keep MSVC6 happy, earlier in the codebase a preceding programmer had coded:
class MyExceptionClass;
Which did indeed keep MSVC6 happy. MSVC9 asked what I’d done with the copy constructor. Well, dear friend: it’s in the header file that wasn’t included. Delete the forward declaration, add one #include and we are cooking with gas.

If we actually give the compiler the definition of the class being caught it can generate code rather than misleading error messages. Why the error message wasn’t: “No definition for this class” is beyond me.

Typical URLs and how to shorten them

Some web sites have very well designed, brief, URLs.

But some have URLs that are way too long. And you don’t always want to be putting them through TinyURL.

Here’s how some of them can be shortened if sending them via email (when they might break when text wraps) or in print.

Anything that’s not bold can be chopped out. And remember when putting it in print, drop the http:// — it’s not necessary to key in, and only slows people down. The same is usually true for the www — though I’m in two minds about that. For publicity etc, it sometimes helps to jog people’s minds that we’re talking about web addresses.

Amazon — it’s the ISBN or other identifier which is critical here
http://www.amazon.com/Doctor-Who-Complete-David-Tennant/dp/B000UVV2GA/ref=pd_bbs_sr_1?ie=UTF8&s=dvd&qid=1212896155&sr=8-1

YouTube — remove the country, and any extraneous arguments such as “Featured”
http://uk.youtube.com/watch?v=-BOYAl0F6xs&feature=user

The Age (and other Fairfax sites) — remove the headline text. (This works for their older articles/older URLs too.)
http://www.theage.com.au/national/clouds-loom-as-oil-price-soars-and-petrol-hits-170-20080607-2n9n.html?page=-1

Google Maps — the co-ordinates and zoom quotient (or whatever it’s called) matter the most. Though if you’re trying to specifically point out an address, you’ll need to leave the query in.
http://maps.google.com.au/maps?f=q&q=247+flinders+lane,+melbourne,+vic,+AU&ie=UTF8&ll=-37.813751,144.964621&spn=0.011188,0.018196&z=16&iwloc=addr

Realestateview — gets really messy depending on how you find the property. Some of the arguments tell it what navigation options to show, but when it all comes down to it, it’s the OID which is the critical argument. Mind you, leaving the “rev=on” stuff gives you the area map by default, so better to leave that on if emailing.
http://www.realestateview.com.au/cgi-bin/view.pl?OID=1136439&rev=on&s=102294592&Sub=bentleigh&BeL=0&BeH=9999&PrL=0&PrH=99999&Surr=&IKW=bolinda&PT=hou&PT=uni&PT=tow&PT=stu&PT=lan&PT=dev&PT=inv&PT=ter&PT=vil&PT=sem&PT=dup&PT=pen&PT=wac&PT=hol&PT=rta&PT=alp&PT=car&PT=bof&PTr=&CS=VIC&OrderBy=listed&OrderStr=&Con=S&SearchPage=/buy/residential/melbourne.shtml&Bkmk=_&OFI=&OFIDays=&BS=10&Thu=&Qui=n

BBC News — these aren’t overly long, but can still be shortened.
http://news.bbc.co.uk/2/hi/south_asia/7442323.stm

MS KB — all kinds of different versions of their URLs fly around the place, though a lot of their new links use the most sensible, concise version.
One of many of the old style: http://support.microsoft.com/default.aspx?scid=kb;en-us;Q917925
Better: http://support.microsoft.com/kb/917925
(Of course to many geeks, just say Q917925.)

For other sites? No doubt people will have their suggestions.

For myself, when I’m sending a URL to someone, and I have the time, I tend to muck about and remove what look like the extraneous parameters and see what still works. Mind you, some sites don’t work very well for this — Dick Smith (dse.com.au) for example, relies on some kind of weird-arse session parameter, so it’s best to use their own “email this link” feature.

And always check it before you send it.

Weird bug

Let’s say, for example, that a system supplies you the time of some event in UTC, you convert it to local and shove the date/time up on the display. Say, for argument’s sake, you also include the Day Of Week, ending up with a format of DD/MM, DOW HH:MM. Everything looks fine, until someone notices that the Day Of Week is wrong. The 28th of May is a Wednesday, not a Thursday.

What happened?

The date conversion routine that generates the DOW string does a bunch of odd stuff, but seems to work correctly; it certainly works in other parts of the code, and generates the right string there.

WTF?

The UTC time seemed to be converted to local time twice, but that wasn’t the culprit; surprisingly, no-one is killed in an explosion of silicon splinters when that code is double-executed. Whatever.

Could it be that the system supplying you the time of that event in UTC is off by a year? One year into the future. That would give you that behaviour.

Check it.

Freebie database dev tools

I used TOAD for a while for browsing around Oracle databases. Good stuff, though the free version is a bit limited, and Quest Software, who sell it, have the annoying thing of not selling it retail, but negotiating licence prices, making purchasing it a long drawn-out process.

Turns out there’s now an alternative: Oracle’s SQL Developer, which does a pretty good job at doing the same thing, and it’s free. It runs on Windows, Mac and Linux, and it uses Java and the JDBC client and apparently doesn’t need the humungous Oracle client software (though I haven’t actually verified that yet).

(Tony found the SQL Server equivalent some time ago, though it’s moved slightly. MySQL? MySQL GUI Tools.)

The importance of filenames

When I’m putting PDFs (or any kind of downloadable file) on web sites, I’m always sure to publish them with a filename which is meaningful, including some indication of the web site it came from. ‘Cos if the average user is like me, they save PDFs etc all over the place, and some kind of identifier makes it heaps easier to find the file later.

Car rental web sites again

(Previous rant)

Trying to hire a car in Adelaide for a single day, a Sunday. I’d really rather not have it overnight, so I don’t have to park it somewhere. As it turns out, it looks like none of the car rental places are open beyond 4pm on Sundays, but the problem with their web sites is how they tell you this:

Budget: “Your return time is after the location’s closing time.” — So what’s the location’s closing time? Surely you’re not going to make me guess? Wait, check the Locations page. OK, found the list of nearby branches. Click on the chosen one. No opening hours shown. … (sometime later) … Wait, if I mouse over the location on the map, it tells me the opening hours! Not that there’s any prompt to the user to say that’s how it works.

Hertz: “Return Date or Time – This location is closed at the time indicated. Please adjust your return or select an alternate location. [DE357]” — But at least there on the screen it’s showing me what the times are.

Thrifty: “Your rental details could not be processed because the return time is outside the opening hours for the return location. The return time has been changed to be the location closing time.” — Better, but why didn’t you tell me the opening hours when I was checking what your locations are?

Europcar: “Your Return Time could not be accepted because the operating hours for our Adelaide City Branch on 10/28/2007 7:00:00 PM are between the hours of 09:00 and 16:00.” Bravo!

Avis: tells as I click through where the branch is and what its hours are, but ignores it for another screen or two. But when it does notice, it eventually tells me everything I need to know in one message: “The Return Location selected is closed at the time requested. The Return Location operates from Sun 08:00AM-02:00PM; Mon-Fri 07:30AM-06:00PM; Sat 08:00AM-02:00PM. The Return Location Service may be available after hours. Please call this number (61) 08-84105727 to contact this location for further details.” — Bravo Avis, you win.

Of course, the aggregator sites like VroomVroomVroom don’t do too well out of this either. Oh well. SimilarlyDriveNow.

I ended up booking with Avis (closest to the hotel) via DriveNow (the best mix of good prices and nice web interface). And when I discovered (after placing the booking) that I’d entered the wrong dates, I went to avis.com and changed them. Neato.