timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Mon Feb 17, 2014 11:41 pm

timrowledge wrote:Special thanks for finding a test case for the block-drag-lockup. I'd seen it a couple of times but never spotted the pattern. I should be able to work out what is messing up now. The immediate problem is that the block you are dragging things it is its owner and you get an inconvenient loop.
OK, I finally found the cause of this. Debugging very-deep system code can be such fun; terribly prone to Heisenbugs and total crashes. What joy... and headaches.
Anyway, problem solved now, will be released next time.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Mon Feb 17, 2014 11:43 pm

ghp wrote:testcase: snap in range
have a block on stage
drag another block below. When in snap range, a white bar is appearing. Drop block just outside snap range. The distance to block is a little bit larger then in scratch 1.4. prox 15 percent. I noticed and confirmed by measuring.
I can't find any difference here. None of the threshold values during the search for attachment points have changed at all. If you can spare a moment to post a couple of comparative pictures it might explain things to me.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Tue Feb 18, 2014 12:04 am

ghp wrote:Let me add another minor observation:
2014-02-12-213304_1920x1080_scrot.png
- fixed ; a simple layout SNAFU
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Wed Feb 19, 2014 11:14 pm

Hello Tim,

you are right: pixels are the same. Proof:
blocks.png
blocks.png (5.68 KiB) Viewed 3800 times
Right side: scratch_rasp_1.4, left is scratch_rasp_rel3.

Sorry for the late response. My daytime job keeps me busy, and at night I optimized some of the kids scratch projects.
Regards,
Gerhard

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Wed Feb 26, 2014 12:59 am

ghp wrote:Project created on windows with 1.4.
Moved to raspberry, runs in rasp-scratch 1.4 and rasp-scratch-rel3.
opened in rasp-scratch-1.4, saved, open in rasp-scratch-rel3 and 'reading' stays there for ever or till my next power bill is arriving or I 'ctrl-C' NuScratch3.sh.
OK, this is a side-effect of my mistake a while ago with the olScratch release where the remote connections were left enabled; saving the project includes that status and the new system doesn't yet hookup properly. So you get a nasty lockup. Sigh; more bugs to fix...

If you carefully turn off the remote connections in olScratch *then* save the project and load it into rely (coming soon to a Pi near you) it works ok.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 1:34 am

OK, for those of you paying attention, there is the next step of making this stuff work better available at
https://www.dropbox.com/s/t7rh1n09xmaz8 ... ratch4.tgz
The dragging of CBlocks now works, assorted pixel-gribleys are gone and the popup menu actions all (so far as I can tell!) work.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 2:44 pm

Hello,

speed of editing is really nice now.

Unfortunately there are a few flaws.

Test: C-Blocks: build a script (two blocls) add hat (green flag). Drag 'forever around this.
I found this working, but in some cases some pixels remained undeleted on the screen.
graphics_cleanup.png
graphics_cleanup.png (5.53 KiB) Viewed 3639 times
This happens when you work at the top edge of script tab. Drag two blocks close to top edge, let hat snap in touching or almost touching top edge.
Does not happen 'in the middle'.

Test: drag script to somewhere on IDE. Unless its in palette, it moves to script tab.
Works for scripts with hat plus blocks, hat only or only blocks.
If you drag to to tab headers (drag to 'costume'-text or 'sound'), it stays there. Can drag it away from there into script tab.
hat_on_top.png
hat_on_top.png (5 KiB) Viewed 3639 times

Test: undelete. Undelete level is one deep in scratch.
delete a block, undelete works
undelete again, nothing should happen. There is really no action then, the system is frozen (even ssh) and my action is reboot.

Test: comments for blocks
working

to be continued
Last edited by ghp on Fri Feb 28, 2014 3:30 pm, edited 2 times in total.

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 3:26 pm

continued

Test: place blocks somewhere. Drag a script to snap in top of (here red blocks). When snapped in, the red blocks are not correctly displayed. When dragging away the blocks from the hats, the red blocks reappear.
prepare_to_snap.png
prepare_to_snap.png (6.93 KiB) Viewed 3631 times
and_snapped.png
and_snapped.png (5.15 KiB) Viewed 3631 times
The screenshots are taken with blocks close to top of script area. But happens wherever the snippets are located.

Test: remote sensor connections
When enabled, the system freezes. Regardless whether a client is connecting or not.

Test: performance: script for cat.

Code: Select all

repeat forever
    turn 15 degrees
    move 20 steps
In scratch 1.4, this increases system load, not reaching 100 %. in rel4, this drives system load to 100%
In scratch 1.4, 1000 movements take 45 secs
In scratch rel 4, 1000 movements take 42 secs
These rotation/moving stuff is challenging graphics.

Test: performance: glide in a square.
Performance is almost the same for scratch 1.4 and rel 4. But if the sprite is rotated, in scratch 1.4 performace is same and in rel4 it increases. Guess the rotation is performed each time in rel4, and in 1.4 its cached ?
move.gif
move.gif (5.09 KiB) Viewed 3631 times
Test: middle click script blocks, results in exceptions.

Summary: great improvements. If one of the next things would be the sensor connections, I could test my scratchClient and challenging sockets.

Regards

Gerhard

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 6:18 pm

Wow! Thanks again for testing so thoroughly.

I'll add bug reports for all those to my list.

I think I can see how the "Test: drag script to somewhere on IDE. Unless its in palette, it moves to script tab." might happen. The "Test: C-Blocks: build a script (two blocls) add hat (green flag). Drag 'forever around this." mystifies me right now, but that's the joy of bugs. Odds are it's a sign-inversion or other dumb-thumb error in calculating the damage rectangle.

The undelete buffer is indeed only one layer deep (I had to work on an infinite-undo system a few years ago for a document processor - it was a nightmare. They're *really* complex) and ought to do nothing gracefully when empty. Ah; a thought- when the undelete 'does nothing' it makes a beep of some sort. This might be the sound system driver causing the problem. If you edit the NuScratch.sh script to change -vm-sound-alsa to -vm-sound-null it might not happen, indicating the sound is at fault. I've been having much trouble with sound recently, though building a new SD card appears to have helped for now.

"Test: place blocks somewhere. Drag a script to snap in top of (here red blocks)." - that almost seems like the blocks didn't actually get added to the script. Hmm.

Remote sensors - I've been working on the socket stuff that handles this; you won't be terribly surprised to read that over ten years the socket layer code has been changed to improve it. This means that old code which used the old socket layer needs finding, understanding, fixing and testing. I *think* I've got it better, if not correct yet.

"Test: middle click script blocks, results in exceptions." - this is to get the popup menus? Or do you have the right button mapped to popup menus? (I know a lot of people do. It's one of the great moral failings of the modern world. Everyone knows that the left button should do selection actions, middle button pops up context sensitive menus and the right does meta actions. It's just How It Should Be). If you're not expecting a popup menu, then that button ought to be bringing up the 'Halo' with a load of interesting buttons around the edge. Click anywhere and it should go away; for the production system that will be turned off but it's a useful way to get to lots of debugging info right now. I'm unable to get any exceptions from clicking any button on script right now. Anybody else?

Performance - yes, some things are oddly slower as well as some being nicely faster. Part of this is likely due to the changes in the UI framework. For example, it is possible to direct all the graphics to external systems (think something along the lines of screen sharing in Skype/Facetime, or shared worlds like mine craft etc) and that means more code. I think the deep layout code used for the reporter blocks might be costing a lot of time too. And the input event stuff is causing issues - we still miss keyboard events a lot, which seems to be down to X right now. It isn't a problem that I've ever seen on RISC OS on the same Pi, for example.

Thanks again for excellent testing! I'll get a next step out as soon as I can fix some of these problems, and hopefully get the i18n stuff squared away.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 6:30 pm

For the "middle click on block":
Wanted to 'print script as image', which is usually right mouse. Currently, I need to use middle mouse for all this context things.
And my description is not precise enough: 'context click' on a hat.
click_hat.gif
click_hat.gif (25.76 KiB) Viewed 3606 times
For the performance:
both scratch versions 1.4 and rel4 have been working on same machine. So differences in X-Performance could not be the reason.

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 9:13 pm

Ooh; just found out what you mean about the exception when expecting a popup menu. Excellent. I can fix it now...

Aaaand, stupid typo. Nice.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Feb 28, 2014 9:30 pm

Test: performance: glide in a square.
A nice test. I stretched it a tiny bit to 'think' the time at the end and then duplicated to have one that doesn't rotate the cat. On my Pi the times are almost identical (40.78 & 40.81 secs) which is interesting. It's also within a gnat's eye blink of my iMac at 40.62sec, so the new VMs timer/ticker handling is working nicely. (the old vm has an.... interesting... view of what time it is on unix. 'Not exactly consistent' is a polite way of putting it)
Attachments
squaredancingcat.sb.zip
zipped because you can't attach .sb files. sigh.
(41.86 KiB) Downloaded 279 times
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Sat Mar 01, 2014 3:19 am

Test: C-Blocks: build a script (two blocls) add hat (green flag). Drag 'forever around this.
I found this working, but in some cases some pixels remained undeleted on the screen.
Rats. This is not going to be fun to track down. :-(

It's related to the other dragging bugs with CBlocks but not directly, I think. On *some* occasions some state is getting mangled so that the damage rectangle reported when repairing screen areas is not *quite* correct. It's a mildly persistent bit of state, since the effect lasts until the clock size changes as a result of the dragging. It seems to be quite hard to provoke, so that's going to make life even more fun. Oh well, that's what we get paid for...

[edit] Interesting - I can actually cause the effect in original Scratch, just occasionally.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Mar 07, 2014 1:26 am

timrowledge wrote: Rats. This is not going to be fun to track down. :-(
Oh my, that was painful. Looks like I've fixed it at last.
That seems to be all the issue bhp noted solved, so on to internationalisation mixups!
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Wed Mar 19, 2014 8:30 pm

By the time most of you get to see this the release5 tgz file should be available on DropBox at -

(Edited to update URL - https://www.dropbox.com/s/oq43vfuv9gfzb ... ratch5.tgz is no longer valid,
https://copy.com/F01OLiWrf4GCP6Kf is the replacement location)

Pretty much all the problems reported at the beginning of the month are fixed (except for some jpg loading) and the block dragging has been massively rewritten to work more cleanly.

The internationalisation (that's why we nearly always abbreviate it to i18n) has been turned on and seems to work within the limits of the font rendering capability we currently have. I would warn you that it takes a *long* time for the menu of available languages to appear when you click on the globe/world icon. Pretty much any language that use a latin-1 like character set should loo ok and if you see lots of ??????? then you've chosen a language with trickier characters. So far as I can tell even right-to-left languages work layout-wise.

The socket based broadcasts ought to work now, as should the Mesh stuff.

The quarter-size and presentation display modes still don't behave well. Soon....
And the sound system is still causing problems at the very low (well below any stuff I can change) occasionally. I see reports of similar issues with other applications.

This should be the last alpha release; unless any catastrophic problems cause us to have to retreat we'll be moving to beta - i.e. everything is there even if it doesn't work quite right. After that it's back to making everything much faster...
Last edited by timrowledge on Mon Apr 21, 2014 6:48 pm, edited 2 times in total.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Thu Mar 20, 2014 6:14 am

Hello Tim,
it is fun to work with this release. Unfortunately I am short in time for the next days.
- Timer display not correctly, still having a lot of trailing digits.
- Variable display also a problem. I displayed a variable with the value (1/3), and result was 0.30000000001. I checked with compare statements, and internally it is 0.33333.
- Sound still having trouble. After first sound output with this release it froze my system.
- in legacy scratch, the gotoXY is prefilled with (0,0), in new release it is something like (548,0)
Regards,
Gerhard

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Thu Mar 20, 2014 9:34 pm

ghp wrote:it is fun to work with this release.
Glad to hear it!
ghp wrote: - Timer display not correctly, still having a lot of trailing digits.
- Variable display also a problem. I displayed a variable with the value (1/3), and result was 0.30000000001. I checked with compare statements, and internally it is 0.33333.
Wow; that's very strange. The value is actually held as a proper fraction (did you know Smalltalk does actual proper fraction arithmetic? I can't help thinking there must be a neat tutorial/lesson that could be done with that facility) and clearly something very odd is happening to get such a wrong display string.
ghp wrote: - Sound still having trouble. After first sound output with this release it froze my system.
It drives me nuts. So far I can't find anything at all sound related that has changed between the very old VM code and the latest. It works on many other unix machines. The Smalltalk code works on many other operating systems, including RISC OS on the Pi. I see reports of other ALSA sound using software causing issues on the Pi and indeed on other machines.
ghp wrote: - in legacy scratch, the gotoXY is prefilled with (0,0), in new release it is something like (548,0)
Typically it is set to the position of the sprite that the 'motion' category tiles are attached to. 548 is a very strange number to see there.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Mar 21, 2014 2:29 am

ghp wrote:- Variable display also a problem. I displayed a variable with the value (1/3), and result was 0.30000000001. I checked with compare statements, and internally it is 0.33333.
Ah, found it. Formatting numbers is actually quite complex... but now it seems to be correct and fits the result into the watcher.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Sun Mar 23, 2014 2:42 pm

Hello,

again some observations.

- Have two objects. Single click on object (in lower right pane) to switch to other view.
Takes quite a long time to switch. 20 secs in some cases. Double click changes immediately.

- contrast still low.
contrast.png
contrast.png (8.54 KiB) Viewed 3244 times
Texts are bold in 'raspi-1.4-scratch' ?

- project preview in open dialog.
Open dialog does not display preview image.

- tooltips are 'bubble style', in 'raspi-1.4-scratch' it is rectangular.
- no tooltips on duplicate, delete, enlarge, shrink buttons (top of stage), buttons in graphics editor.
- open a sprite in graphics editor, use the buttons (turn, toggle left right ...) each after the other. The single buttons keep the 'pressed' look and action is not the one intended.
- set sprite's rotation point not working. I added a long zigzag-line and the rotation point to the extreme. Market this by an X. On stage, rotation is not around the rotation point, but somewhere else. Same with goto xy, should set the 'rotation point'.

Worked a little bit with remote sensor protocol, and it works. Also the "scratch-startclicked" event. Btw, are there other 'scratch-NNN'-Events recognized ?
- opening a project which did have 'remote sensor enabled' do not open with this setting enabled (although I do not like dialog popup, I miss it...).

Regards,
Gerhard

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Sun Mar 23, 2014 5:47 pm

ghp wrote:- Have two objects. Single click on object (in lower right pane) to switch to other view.
Takes quite a long time to switch. 20 secs in some cases. Double click changes immediately.
Hmm. This might be related to the open dialog/preview problem, since it seems like it should be mouseDown/move/hold/up event related. There a re several places where I still need to sort that stuff out better.
ghp wrote:- contrast still low.
Font choices need fiddling with.
ghp wrote:- project preview in open dialog.
Open dialog does not display preview image.
Oddly enough I fixed this up just yesterday. I noticed that using the cursor keys triggered reading the preview and that reminded me of the mouseXXXX event changes. I had much fun working out what was supposed to happen when browsing for sounds to import - did anyone notice that the sound sample will try to play when you click on it and stop if you click again?
ghp wrote:- tooltips are 'bubble style', in 'raspi-1.4-scratch' it is rectangular.
I think you mean the popup menus here; yes, the corner rounding needs to be turned on. I have it set to not do that in my working systems because I really, really, dislike it. Forgot to reset it.
ghp wrote:- no tooltips on duplicate, delete, enlarge, shrink buttons (top of stage), buttons in graphics editor.
Good catch; I hadn't noticed that.
ghp wrote:- open a sprite in graphics editor, use the buttons (turn, toggle left right ...) each after the other. The single buttons keep the 'pressed' look and action is not the one intended.
Ditto.
ghp wrote:- set sprite's rotation point not working. I added a long zigzag-line and the rotation point to the extreme. Market this by an X. On stage, rotation is not around the rotation point, but somewhere else. Same with goto xy, should set the 'rotation point'.
Err. Works ok for me.We'll have to more carefully compare exactly what we are doing.
ghp wrote:Worked a little bit with remote sensor protocol, and it works. Also the "scratch-startclicked" event. Btw, are there other 'scratch-NNN'-Events recognized ?
Hooray! Events and which ones are handled has not been altered at all, so if something isn't working that used to, it's a bug.
ghp wrote:- opening a project which did have 'remote sensor enabled' do not open with this setting enabled (although I do not like dialog popup, I miss it...).
Not sure how this can be; one of the bugs we had a little while ago was caused partly by the remote señor ( I swear I typed 'sensor' there) enabled state being carried across properly. Wonder what I broke?

Thanks again for taking the time to document your testing.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Sun Mar 23, 2014 6:14 pm

set sprite's rotation point not working.
Hello, I have to take this back. Build several examples and works in all scenario.
Regards, Gerhard

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Mon Mar 24, 2014 10:22 pm

timrowledge wrote:
ghp wrote:
ghp wrote:- opening a project which did have 'remote sensor enabled' do not open with this setting enabled (although I do not like dialog popup, I miss it...).
Not sure how this can be; one of the bugs we had a little while ago was caused partly by the remote señor ( I swear I typed 'sensor' there) enabled state being carried across properly. Wonder what I broke?.
I simply can't replicate this right now. Very odd. Perhaps you could upload a project that you think should have the remote sensing 'on' but that loads without it on your machine?
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Tue Mar 25, 2014 1:04 am

timrowledge wrote:
ghp wrote:- tooltips are 'bubble style', in 'raspi-1.4-scratch' it is rectangular.
I think you mean the popup menus here; yes, the corner rounding needs to be turned on. I have it set to not do that in my working systems because I really, really, dislike it. Forgot to reset it.
ghp wrote:- no tooltips on duplicate, delete, enlarge, shrink buttons (top of stage), buttons in graphics editor.
OK, found the cause of the missing tooltips and realised what you meant about the rectangular shape. I've got them working now.
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

ghp
Posts: 1395
Joined: Wed Jun 12, 2013 12:41 pm
Location: Stuttgart Germany
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Thu Apr 17, 2014 3:37 pm

Hello,

found some code which does execute in raspbian-scratch-1.4, but not in nuScratch5.

Purpose of this code is to receive some semicolon-separated values in a variable, separate these values, calculate scaling factor and display as a chart on screen.
I have extended my atmel-adc-firmware to record timestamps on the signal edges of an ir-control-receiver, and want to display these in scratch (why? because its possible).
The code does not work on the "if ; == letter(i, daten)". Obviously in new scratch, this never returns a ';'. In legacy scratch, it works.

Regards,
Gerhard
Attachments
irc_display.zip
(42.46 KiB) Downloaded 257 times

timrowledge
Posts: 1272
Joined: Mon Oct 29, 2012 8:12 pm
Location: Vancouver Island
Contact: Website

Re: A new Scratch; pre-alpha-this-will-hurt release

Fri Apr 18, 2014 3:33 am

ghp wrote:The code does not work on the "if ; == letter(i, daten)". Obviously in new scratch, this never returns a ';'. In legacy scratch, it works.
OK - I'll take a look at that.

Sometime pretty soon we'll be stepping over the line to a proper 'Beta Release' - which I define as "everything we can think of is actually in there, at least sort of works, and it doesn't crash too often". There will, of course, still be bugs. It's software after all; one of my favourite siglines says -

"The software is only finished when the last customer dies"
Making Smalltalk on ARM since 1986; making your Scratch better since 2012

Return to “Scratch”