When I look at Scratch, I see something ‘different’.
I’d like persuade you to my viewpoint by going through what I see are the Good Points.
I’ll be up-front here and state my target audience is the ‘noble corporate toiler’. It has hopefully been a while since you were introduced / subjected to a computer based teaching tool. Scratch is a tool for implementing computing projects based upon many years of thought, and it is well worth taking a look.
Just what is Scratch?
It’s a visual programming language, consisting of composable code blocks that form the elements of a program. It is available as a web application online, and implements a visual stage as the program output.
There is also an offline version now, which largely matches the basic capabilities of the web version, omitting online-only features.
A whistle stop tour
Scratch operates on a visual stage of 480 x 360 virtual pixels, and can be viewed at varying dpi – Good Point 1: fixed stage extent. The stage is a sprite [Wikipedia-2] and can be given script code, and any number of sprites can be placed on the stage to compose a visual scene. The sprites have a scripts container, which owns aggregations of code blocks that can interact with stage changes, key and mouse events and manipulate the sprites’ properties. The sprites have a concept of their direction and can be made to orient in any direction, move and ‘bounce’ within the stage area automatically. Finally, the large array of primitive code blocks available for the scripts can react to events and messages and with the sprites, and allow the programmer to construct a range of different behaviours and visual experiences.
Figure 1 is a grab of a minimal project put together from an empty project and the built in resources.
The entire (basic) environment is visual, with editors for all aspects of the sprites and code blocks. The stage is manipulated at both ‘run time’ and ‘design time’ in the same way, which is Good Point 2: persistence of stage state. This allows fine positioning by eye, avoiding too many demands on mental arithmetic for the initial learning stages.
Image and audio resources are embedded into to the project and are available in a standard library – Good Point 3: built in resources – which can be expanded by assets uploaded to the project. All of these can be copied, and modified once embedded.
Running the project
There is a Green button to start running, and a Red button to stop running. What running means can be deceptively simple – the simplest event is ‘when flag clicked’, which can have some code blocks attached underneath – this could perform initialisation or start a processing loop of some kind. You can have as many as you like.
Debugging is an interesting proposition for the target audience.
- Adults (example) @garybernhardt: “I’m in a super good mood. Everything is still broken, but now it’s funny instead of making me mad.”
- Years 3–8: Now, I’ve seen their projects, and for some let me assure you: ‘Everything is broken’, but they’re not fazed.
There is a learning curve on the road to learning to debug your project, and there are some useful built-ins, like message and comment notes. The coolest of which, though, is: Good Point 4: code blocks can be modified at runtime in the designer. All versions allow blocks to be dragged in, then dragged out while the project is running.
Can your tool do that?
Figure 2 is a grab of the code blocks being inspected while the code is running and Figure 3 is a grab of a chunk of code blocks after being pulled out of their container block while the code is running.
Did I mention you can do this while the code is running?
You can do this while just using the mouse – the selection extends naturally from your selected block to the end of the enclosing container block – making Good Point 5: selection of blocks has some subtle but powerful affordances. Making use of Good Point 6: the positioning of code blocks in the scripts area can be arbitrary, you can park a small chunk of logic within the visual context of where it was running to reduce the cognitive load.
This works particularly well on an interactive whiteboard to show the effect of a chunk of code blocks.
Even more impressively: the Old Scratch 1.x desktop version, has a similar single step, which illuminates the currently running block! Clearly this was too mind-blowing and is no longer available in the current version.
Why should I be interested in Scratch?
A good question. Here are some topics to persuade you.
It’s coming to a ‘classroom near you’
Good Point 7: ‘This Thing is Happening’ Note also that Scratch at this moment is on the way to being taught in the UK in primary and secondary schools – that’s years 3–8! (around 7 to 12 years old). [gov.uk]
The Scratch heritage
The Scratch environment is inspired by the work of Seymour Papert [Papert], epitomised in the book Mindstorms [Papert93]. Listing the whole corpus of the back story is out of scope, so this is going to be left for the interested reader.
You can explore for yourself by visiting the website [Scratch-2], making use of Good Point 8: it has a single entry search box. Having found a project of interest, you can dive straight in and take a look using Good Point 9: Scratch has a ‘See Inside’ button. If you like what you find you can then simply fork it using Good Point 10: Scratch has a ‘Remix’ button That new project is now available for you to modify, debug and run in any way you see fit.
Now, although it ruins the nice 3-part list rhetorical flourish, let me just mention here Good Point 11: Scratch lets you edit before login and then allows you to login and save if you wish.
Question: how many paid for services that we use for generating wealth can say the same? For the classroom environment, when dealing with the early years where it is a stretch to require detailed forward planning, this is a Good Thing.
What can be done in Scratch?
I won’t promote individual projects, and further I will assert that I don’t need to. Instead let’s rely upon Good Point 8: it has a single entry search box and I can list some jump off points:
- Yorkshire to English dictionary [Scratch-Dictionary]
- Innumerable RPG type things about cat clans [Scratch-Cats]
no, me neither
- Space Invaders [Scratch-Invaders]
- Pacman [Scratch-Pacman]
Who are the users?
Although it seems to be primarily educationalists and UK scholastic years 3–8 (at least), there are 10s of millions of projects and millions of users, leading to Good Point 12: a cast of thousands. [Scratch-Stats1] [Scratch-Stats2]
A list of additional Good Points
There’s a raft of features: Good Points 13–21:
Logo-like Sprite Primitives*, Code Blocks*, Message Passing, Composable Mathematical Operations*, Sprite Cloning, Entry Prompts, Text Messages, Pen operations and Sound. The annotated items (*) are all shown illustrated in Figure 2 in one way. The items listed are in my personal order of appreciation. A feature worth mentioning is Message Passing, which allows broadcasts of user defined messages to all items on a stage.
Supported development styles
So, the possibilities are limitless, but there is a definite set of approaches that practitioners will work through.
- Basic: sprites, backdrops, costumes
Coupled with setting some properties, the bounce/direction/ touching colour capabilities allow treating the visual stage as a mini engine with some rudimentary support for using the stage as a dressed ‘set’.
- Advanced: explicit positioning/drawing
To move onto to more dynamic generation of content, it is possible to use explicit positioning, and the use of pens to craft arbitrary shapes.
This is the type of benefit mentioned in the kind of thing Brett Victor has talked about [Victor] – the good stuff, in my opinion, started around 10:00. A brief summary is that he demonstrates a ‘Braid’-like game [Braid] in a live editor, which has the ability to adjust variables using sliders at run time and to record and overlay timelines of the program state. The gist of the open question raised was “What could you do , if you could only visualise the outcomes?”
I still think about the impression that talk had on me. It’s not quite possible to reproduce that talk right now in Scratch, but the immediacy of the run–change–run loop is still very forceful.
Why haven’t I mentioned entirely free to use yet? If you care about diversity and granting access to self-improvement to all then this is huge.
From Good Point 10: Scratch has a ‘Remix’ button
- It exists and works – if you like something and want to have a try at improving it, you just press the button
- Is this due to Good Point 22: Scratch has No Merge Action (for you corporate warriors)?
This point maybe bears some examination – there are no libraries in basic Scratch and no code sharing. In order to add some code inspired by code from elsewhere, one has to understand what to splice in to a project, and actually do it. Because there are two ways to generate Scratch projects – remix and create new – there is at least some pseudo Darwininan process that improves the fitness of remixes, while the ‘gene pool’ gets a steady stream of new projects.
From Good Point 23: Scratch has a ‘public’ checkbox, almost nothing could be simpler – you just click the ‘public’ checkbox.
From Good Point 24: Scratch seems to be incredibly good natured, is it because they’re all under 12? (or over 30?) Well, there are The Rules [Scratch-3] and there are the banned topics [Scratch-4]. There even seems to be a Scratch version of Herobrine [Scratch-Herobrine].
Don’t worry: it’s not like ‘coding’
It is important to point out that ‘just coding’ is not the point – it’s engagement with the environment, using creative approaches and problem solving that are the real end games here. [code.org]
Suitability for classroom and distributed teaching
Good Point 25: Scratch supports multiple sessions on the same account and concurrent work on different projects This is one setup I can advocate: given each project has a thumbnail in the account’s list of projects: Good Point 26: all projects have a thumbnail makes it very easy for a supervisor to see what individuals are up to. This also allows the supervisor to investigate and debug a project remotely if needs be.
Teachable moments 1: big chunks of blocks vs message passing
When did you, dear reader, love to learn message passing as a style? In the Scratch environment, it seems to be a few months from a standing start. Very quickly, the users seem to grasp that giant chunks of nested blocks and variables can be replaced by sending the appropriate message to be handled by a smaller chunk of code blocks.
Teachable moments 2: Space Invaders vs Pacman
This seems to be one of challenges that sorts the population. Caveat: this is from my purely personal sampling. I have found the Space Invaders clones tend to be very high quality, whereas for the ‘deceptively simple’ Pacman there tends to be a raft of issues that challenge the users Good Point 27: some tasks lend themselves to ‘teachable moments’. My suspicion is that there are a couple of required complex concepts in a maze game that Scratch neither delivers in its default toolbox, nor allows to be easily synthesised. This might be a fruitful area to extend the sprite’s capabilities:
Confession time: for the second time (approx 30+ years after the first attempt) I am re-implementing Pacman and I’m finding that talking about the incomplete (broken?) versions I created on the way can be used for some ‘teachable moments’, for example:
- various ‘baby step’ projects that show the small adjustments to the code blocks that will implement the stages of a ghost
- moving smoothly between points on a grid
- choosing between N, E, W, S to head towards Pacman/some other target
- turning left instead of reversing direction on the next choice (which delivers the distinctive Pacman ‘ghost patrolling in circles’ behaviour for free)
User straw poll
It hit me that I should get an assessment from the horse’s mouth, as it were. I therefore ran a very unscientific poll of a miniscule sample set of users, and asked them one question.
The results for answers to “What’s the best thing about Scratch?” are:
|You can draw costumes||1|
|You can use your imagination||2|
|You can make games||4|
|It’s easy to share||1|
|It’s satisfying to finish things||1|
So, making games wins it, but the ability to express yourself is clearly highly valued. It’s very encouraging that user agency afforded by working in Scratch is prized over passive consumption.
Can it really all be Lovely In The Garden?
Sounds too good to be true, right?
- Well, it is a social environment with quite young users interacting over the internet running arbitrary code. Security concerns will always be with us, but there as mentioned are rules and admins which (touch wood) seem to work.
- Is it so addictive as to prevent people progressing to (say) HTML and python? This is a concern I have heard expressed: Scratch can seem to easily become the tool of choice to prototype an idea, which is dual edged: it’s a good thing to get an idea captured quickly, but conversely it would be good to stretch oneself and branch out.
- There is a very serious number of games, some of which are quite impressive and diverting...
Conclusion and lessons for ‘grown ups’
So here we come to the punchline: I have joked that some companies could figure out how to write and deploy their business apps using Scratch. That’s not in fact a serious proposition, but I suggest we can see that the secret of the success is all the Good Things, working to make the environment as expressive as possible.
Many operations take one step
A case in point is the debugging example already demonstrated; in fact the users slip naturally into re-initialising or exploring different initial conditions for the stage in a rough manner by dragging the sprites back into position while the project is running. It’s only upon writing the sentence that it has struck me how extraordinary that is. There is an analogue with, for example, the mavens of languages supporting a command line REPL for ‘real work’, and also raw web development. Yet for many commercial systems, many steps exist between us and the prize.
Most things are quick
This multiplies the time savings with the prior point, but more importantly is key to a learning environment that encourages and rewards exploration. This is why coding club ‘classes’ should ideally not need a squad of instructors to oversee and guide work as the users can very well formulate hypotheses and test them themselves in short order.
The ‘Ah, of course’ moments
Hopefully I have shied away from employing too much hyperbole (!) in going over Scratch, and have instead relied upon the Good Points to make my case. In re-reading I see a pattern emerging of key concepts that have been selected and made available in a frictionless manner. This is part of the power of the Scratch approach, as the cognitive load is minimized, driving out distractions from the flow of working with the concepts of the project. For me this is the key takeaway: a real commitment to reducing the superfluous complexity of our tool sets leads to better outcomes.
[Papert93] Papert, Seymour (1993) Mindstorms http://www.goodreads.com/book/show/703532.Mindstorms
Overload Journal #132 - April 2016 + Programming Topics
|Browse in :||
All > Topics > Programming (766)
Any of these categories - All of these categories