Scratch

14God wrote:

I think I'm the only American here who stays up at this hour... 2pm here... oops just blew my cover(location), at least my longitude 

You do mean 2am, right?

bharvey wrote:

14God wrote:

I think I'm the only American here who stays up at this hour... 2pm here... oops just blew my cover(location), at least my longitude 

You do mean 2am, right?

Oops... yeah (shows how tired I was)

At least we have Google. 

MathWizz wrote:

I'm not too far behind... I'm up 'till I finish my research paper. 

I DETEST research papers...

Yeah I like making stuff more then writing about it.

Jens wrote:

if your research paper is about computer science you might always try http://pdos.csail.mit.edu/scigen/ 

ROFLMAO! That's SO incredibly hillarious! I tried it, and it came up with this:

Figure 6: The expected energy of our approach, as a function of block size.

OMG take a look at this:

Figure 2: The effective block size of our system, as a function of response time.

Last edited by rubiks_cube_guy238 (2011-02-20 20:32:02)

For the first time Jens is using the word "alpha" to describe part of today's experimental release!

Jens wrote:

Tonight's build completes and cleans up the generic object interface for sprites to what might be considered alpha level, adding the (non-inheritable) 'name' attribute, and letting users inspect which attributes are inherited (deleted in a child) by selecting that attribute in the ATTRIBUTE block's pulldown menu in the pallette. If the specified property is inherited, the ATTRIBUTE block turns to "ghosted." This build also cleans up some menus and UI details.

What's still missing is dynamic inheritance for costumes and sounds, as well as first class costumes and sounds. I'll explore those in the next experimental releases.

Just wondering, but is there a way to change the options on the <is [] a [number]?>

I would like to say thanks for making a great tool for teaching kids programming. I have been using scratch to teach my kids programming, but I have felt the limitations of what is available in scratch, and I like many of the things BYOB has done.

There is one thing I really miss in both scratch and BYOB. I would love to see first class subroutines. My kids picked up on the idea of using broadcast to call actions on other sprites very quickly. It would be very nice to have a top level 'action' block that acted like a 'script' block (i.e. could receive parameters and would be called per object). It would also be great if there was a way to do polymorphism with these (i.e. you could store an object as a variable, and maybe have a 'tell object run action with parameters'  where the action could be typed in so that runtime discovery is possible). Maybe all that is needed is a block like the sensing 'attribute x of sprite', but where an object could be dropped into the sprite location, and an action name could be typed in. The new changes to make objects first class should make polymorphism possible and first class subroutines would be beautiful.

I did show my kids how to define a variable, with a script in it, and run the script stored in the variable. However they don't tend to use this because broadcast is more natural. I also saw a was to do polymorphism by passing strings to a script block which selects and returns another script using a long list of ifs. Quite ugly to try and teach someone polymorphism in this way. Seems very cumbersome compared to first class subroutines.

llarsen wrote:

There is one thing I really miss in both scratch and BYOB. I would love to see first class subroutines.

Jens and I are both confused about your message because we do have what we call first class subroutines; they are created by THE BLOCK and THE SCRIPT, which are exact ports of lambda for reporters and commands respectively.

Some of the things you want to do seem to me to be satisfied by the ability to make sprite-local blocks.  In particular polymorphism, as I understand it, works perfectly using our inheritance mechanism.

So, could you maybe write some pseudocode for what it is you want to do in BYOB that our latest experimental build doesn't allow?  Thanks.

wmays wrote:

Just wondering, but is there a way to change the options on the <is [] a [number]?>

We're adding object types to the options, but there won't be a way for users to add types.  The idea here is that every possible value should be in exactly one of the categories, and anything you added would overlap with one of the primitive types.

But of course you can define your own predicate blocks to check for any user-defined type you want!

(attribute [instrument]) doesn't work.

Hey guys. I was playing around on Scratch and realised that there was no tile engines. I then made one, which could display a map. This was not fast enough, so I moved to BYOB. Here is my attempt. It is quite fast, but could be improved. What are peoples opionions?