Scratch

BYOB 2.99.018, no big new features, just some bug fixes, including the one about replacing list items.

TheExternal wrote:

http://scratchx.webs.com/
I re-made my website and (i'm still working on it now) for BYOB.

Nice! Thanks for sharing your site here. It's such a blast for us to see how you guys are enthusiastically taking on those "big ideas" behind BYOB3! The Scratch community really is something else. See, that's why we want first class procedures and first class lists in the official Scratch, and not a separate "Scratch4CS" application: We don't want to split this wonderful community, where grown-ups can learn from kids and vice-versa (just look at the age range in this forum thread, it's amazing!).

Okay, so I'm a little bit confused right now, and I don't feel like reading through 31 pages of posts to find the answer 

Is Jens creating this solo, or is it a group of people? As I recall, Jens did create BYOB, and so I was a little confused when there was a new thread made by somebody other than Jens about BYOB. Anybody care to clear this up for me? Thanks 

coolstuff wrote:

Okay, so I'm a little bit confused right now, and I don't feel like reading through 31 pages of posts to find the answer 

Is Jens creating this solo, or is it a group of people? As I recall, Jens did create BYOB, and so I was a little confused when there was a new thread made by somebody other than Jens about BYOB. Anybody care to clear this up for me? Thanks 

Jens is pretty much writing the software, and bharvey is providing input and expertise, as well as documentation and tutorials. Everyone else in this thread has merely been providing examples, suggestions, bug reports, and praise. And maybe a few gripes. 

Last edited by fullmoon (2010-06-15 16:56:48)

coolstuff wrote:

Is Jens creating this solo, or is it a group of people? As I recall, Jens did create BYOB, and so I was a little confused when there was a new thread made by somebody other than Jens about BYOB. Anybody care to clear this up for me? Thanks 

Jens remains the sole programmer of BYOB.  As of BYOB3 I'm participating in the design (mostly this means I taught Jens about first class functions    but also we've been working out a million little user interface issues) and contributing documentation, tutorials, making trouble on forums, etc.

The two of us are the "team" officially, but of course lots of other people have been helping in large and small ways.  Lucario organized this forum thread.  A whole gang have been finding bugs for us.  We have a team at Berkeley developing a new computer science course using BYOB, and they've contributed ideas; in particular, Sally Ahn, a student in human-computer interaction, is doing an HCI evaluation for us, looking systematically for aspects of the design that would make BYOB unnecessarily hard to use.

Hope that helps.

coolstuff wrote:

Okay, thanks bharvey, fullmoon, henley, and johnnydean1! That helps a lot.

lol, we were all to bored not to answer 

xly wrote:

jstout wrote:

Complicated variables/lists to store the three pieces of information and the restore process is very simple.

It's a good exercise to use your method, but in this precise case probably that it is simpler to save position+direction into 3 variables.

Anyway the solution you mention - to use lists as a kind of shared memory - is not a bad one, mainly if you need to handle long lists of variables. In addition a list can store "procedures" ...and everything you like etc It can be an elegant solution to store every variables, or every property of a sprite into a dedicated list.

I'm not sure it's simpler using three variables, or even a list with three items. I'd have to have one custom reporter block to report a list with three items in it (the current x, y, and direction), and a variable to hold the list reported. Then I'd need another custom block with one list parameter which would pick out the three items in the list and do a goto x:y:, then a set direction:.

With the functional approach I only need one custom reporter block which reports the block that, when run, does the goto x:y:, then the set direction:, and a variable to store the block reported. Then to restore the position I just run the contents of the variable. One fewer custom blocks to keep in sync: if I change the order in which x, y, and direction are stored in the list I'd have to remember to change the code in my restore position block. If Jens and Brian give us a way of saving and restoring the pen then I can add that to the functional reporter block and I don't have to change any other code.

I think it's one of the things that you get from functional programming: once you get your head around the central idea, then all sort of things become so much simpler.

John

Last edited by jstout (2010-06-15 19:27:52)

jstout wrote:

With the functional approach I only need one custom reporter block which reports the block that, when run, does the goto x:y:, then the set direction:, and a variable to store the block reported. Then to restore the position I just run the contents of the variable.

With some regret, because I love the elegance of your solution, I think I've thought of a simpler way for most applications:



But your way has the advantage that you can call the restore-state procedure more than once.

bharvey wrote:

The speed-up in tonight's release affects the part of the code that's responsible for suppressing empty argument slots to be automatically filled with zeros by Scratch. Therefore I'm half expecting some new bugs related to zero-filled empty slots and perhaps even involving saving/reading errors to surface. Let's uncover those if they're in there! 

wait... your gonna make the empty slots automatically fill up with a 0? Can you at least keep some of them like they are because if i leave an empty slot and do this:
{ [answer] > [] } I can make the script sense if there's anything in the box at all instead of just numbers. I use this alot.

Yes. It is possible for that to happen. It managed to crash Brian's Mac. 

TheExternal wrote:

wait... your gonna make the empty slots automatically fill up with a 0?

Not on purpose!  Jens is just expecting that some bugs of that kind will turn up.  When you find them, we'll fix them.   

ScratchReallyROCKS wrote:

Well It hasn't crashed my Mac. Go Mac Minis!!

Apparently it's just my particular year's model of MacBook Pro that has this overheating problem.  It hasn't happened since I installed the Fan Control preferences pane that lets me get the fans going faster sooner.