great post. this articulates well what i'm going through right now acutally. if you have any links on user experience ideas/guidelines that would be cool too. i had never heard of the term or the acronym until now

Thanks! I think doing it in real time could be a little disconcerting for the user - I thought a lot about that and in the end decided to copy how Amazon do it

i'm not an expert at it, but doing BDD has been the most helpful thing so far. and by that i mean not only stuff like cucumber, but more importantly the process of first thinking through and writing down what a specific kind of user wants to do in a specific scenario.

so if you were thinking about twoplustwo, two important user types to think about would be regular forum users and new users. what are the tasks a regular user comes to the site to complete? one of them is to see new activity on threads he's participated in. so how do you present info to him to make those tasks easiest? 2p2 has the subscribed threads feature, eg, for the first task. but it's sort of hidden away. if you were redesigning it, you might want to make it more prominent, or even make it the homescreen, etc. now think about a new user. his goals are completely different. you may have to balance conflicting goals for different user types.

the specific answers aren't my point here though -- it's the process of starting with a user and what he wants to do in a specific case. Writing that down with no concept of how you are going to solve it, no thought of UI whatsoever. That's where you need to start.

sort of on this topic, i recently read this article and found it helpful.

very interesting, thanks

In C# when I make an instance of A that inside it has an instance of B which inside that has an instance of C... when I want an event raised by A to be handled by C without A knowing about C i do a.EventName += b.HandleEventName and inside B i do HandleEventName { get{return c.HandleEventName;} }.
First of all is this standard / okay? And my main question is what is the easiest way to do the reverse of this (i.e. event raised by C that gets handled by A without C knowing about) A. I can only think of two ways, one is have B receive an event from C and immediately raise its own event i.e.
And just have A subscribe to B's copyEvent. The other one is something like
Is there any other ways this can be done? Is one of the two above better than the other?

SurveyQuestion needs SurveyIndex field.

Survey doesn't need numQuestions field. (it is a derived value)

Where is your QuestionAnswer table?

Hey guys, it seems like rails, python, and php are standard for REST-based APIs. I want to make different iphone apps that use them to access databases (stuff like recipe APIs, wine APIs, etc). Which of rails/python/php would I do best to spend some time learning? Also, how much of the language would I usually need to know to do this? Thanks.

Just want to say thanks to all those who offered advice with my survey question!

How feasible is xulrunner, chromium whathaveu for a user interface? I feel like redoing a bit of a gameserver and am wondering what user interface would be best.

I'd say webbased LDO but I want multitabling (just pretend I develop a poker server/client). Has anyone experience with say running 4 xulrunner based user interfaces to connect to some server?

The alternative would be Qt I guess (don't care about the recent turbulences it's good enough and will be good enough imo)

i'm looking for a good/easy gui for viewing multiple server logs from multiple machines. Obviously just using the command line is ok but it would be nice to be able to have some software with tabs that can hold a tail -f style log stream.

anything out there like this for mac? i tried to google but didn't have much luck (although could be me being dumb)

Thanks

does a mac terminal not have tabs? or do you mean something else?

I have used multitail before... its highly configurable and allows you to tail logs from any number of different sources simultaneously. Configuration can be a bit of a pain at the start.

http://www.vanheusden.com/multitail/

I finally did it xD

thanks ill chk it out

i was just looking for something a bit more user friendly, like preconfigured login/pass/path to log, easily export/save/search/watch in real time. command line is ok just thought there might be something better

this looks great

Hey guys, Im having a little trouble wrapping my head around something.

Lets assume in C

int n = 1;
float f = 1.0f;

if (n == f) printf("They're equal.\n");
else printf("They're not equal.\n");

When executed, the output is They're equal.

I get the premise of IEEE 754 for converting floats to binary, but my confusion is when you look at the bit representation of both numbers, theyre entirely different, and the == operation compares bits.

Does the compiler cast one to the other for comparison? Im thinking it casts the int n to a float (using IEEE 754?) and then compares the 2. Is this accurate? Obviously you can find the binary representation of n inside the binary of f, but Im not sure how relevant that is to this comparison.

The == operator does not compare bits. The compiler converts the expression to float(n) == f.

The same exact thing happens if you write n + f or n * f or n / f or if you pass n to a function that takes floats or whatever. This really has nothing at all to do with the details of IEEE 754. Indeed, C does not technically specify what binary representation a floating point should take or even how precise it should be or what range should be possible.

I mean.. it would be pretty ridiculous if 1 + 1.0 in C were like 1.0000001 or whatever the hell you'd get if you just added them bitwise without a cast. It would also be completely ridiculous if two numbers with different values and different types were said to be equal because their bitwise representations happened to be equal. C could have been implemented such that a == b is always false if a and b don't have the same type or even 1 == 1.0 is a compiler error, but it's much more convenient to have 1 == 1.0.

This is the way it works in every language I've ever programmed in, including languages in which ints and floats are first class objects, where one might logically expect 1 != 1.0.

Does it specifically cast int to a float, or just the left most variable?

I believe it's always casting the little one to the big one to preserve information.

It always casts up. It'd be pretty ****ty if 1 == 1.1 were true because the 1.1 got casted to an int. Instead, the compiler changes (1 == 1.1) to (float(1) == 1.1). The same thing happens with double a == float b or long a == int b or whatever. C handles them elegantly.

Technically, there are a number of different operators written ==. There's float a == float b, int a == int b, unsigned long long int a == double b, etc etc etc. In each case, C always casts the two variables into the smallest type that it can that contains both of them and doesn't lose any information.

Makes sense, ty

I'm working on a Django web app at the moment that will allow data entry from 8-ball league nights.

I'm not clueless with python, but I am on javascript and their patterns.

I have a form field using jQuery to lookup player names, and i'd like to fill out other form fields after one is completed.

ie.
When I complete FormFieldTeam1Player1 i'd like it to update the name in Match1FieldP1, Match6P1... etc.

I assume this is a javascript thing - what am I Googleing?

Document Object Model (DOM) perhaps.

in the html you do: then in javascript:

I'm getting there (and just discovered Firebug

I'm now stuck with what events to attach too. Looks like jQuery doesn't trigger the normal dom events like "onchange"