Phratch: A port of Scratch on Pharo
phratch.com
phratch.com
Edit: ah, the old Pharo website (accessible through the link at the top), puts front an center "Pharo is a clean, innovative, open-source Smalltalk-inspired environment". OK then. Good job on redesign that loses primary information. /rant
Essentially, the transition cost to these Smalltalk environments (be it Scratch, Croquet, or Pharo) is too high to ever gain critical mass. And that's a shame. Considering how regularly this situation is repeated, I have to ask if it's something inherent to Smalltalk? Or its legacy?
Compared to other Smalltalk implementations, GNU Smalltalk is very different in one aspect: it is pragmatically designed to be a tool rather than an environment.
It has a CLI, loads programs from the OS filesystem, etc.
As I understand it, these environments are self-contained, in a "separate universe" kind of way. No thanks, I want my apps and services to be deployable to the rest of the world, and I want (some of) the rest of the world's apps to be usable in my environment, which isn't the case here.
Sure, Pharo et. al. running in their own kind of VM alongside the rest of the OS is a compromise to address this very problem[1], but like I said in my previous comment, after seeing Scratch, Croquet, and now Pharo rise up and then fizzle in the wider world I ask: what's the point? Where does it fit in today's development world? Is the goal to build a new, clean but segregated OS/environment entirely separate from the existing ecosystem, or is it to provide a development environment that can be used to create applications that are deployable to existing platforms?
If it's the former, I raise my beer, say good luck and on your way. If it's the latter, you've got my interest.
The NextStep platform died because it was the former. Luckily it had two important legacies: it was used to create the first World Wide Web server and client. The server was a network application that could interface with the rest of the world. Then NextStep died. It was then resuscitated when it reverse-acquired MacOS and leveraged that platform to transition to MacOSX, thanks to the vision of one singularly-driven man and tens of millions of dollars. What's Pharo's path?
[1] if the Smalltalk environments run in a VM inside a widespread OS, I wonder if there have been any modern effort in the opposite direction, to have a Smalltalk environment run on bare hardware?
About bare metal language hosting, http://www.openmirage.org/ comes to mind, sure it's not smalltalk or even a lispos but that's still a big step in bare metal abstraction running.
Smalltalk was not designed to be a programming language. It was intended to be the whole system. I think those developers who are treating and developing Smalltalk only as a programming language and a development environment have got it wrong, although it is the pragmatic thing to do. People who are ignorant of Smalltalk simply don't understand how backward contemporary computing its, relative to the way it should be now, 30 - 40 years after the first implementations.
After it emerged, the industry majors simply diverted it way from the original vision of its developers, which was focused on exploring and developing for personal use either recreational or work related, and turned it a development tool costing thousands of dollars.
My own interpretation is that the industry majors stifled it, masking the process as bad business and marketing decisions.
I mean what could be possibly have been wrong with Smalltalk when the IBM's first Java implementations and IDEs were built with Smalltalk, and Sun bought out the company which developed one the fastest Smalltalk implementations to put their developers into creating Java?
I think the current problem with open source Smalltalk is that the developers are spread among different implementations. It would be better if there was one vision and one implementation that they all focused on.
Correct. This is a strength and a weakness. I was enamored to Smalltalk early on (back in 95 when I learned it in a course) for this very reason. But then the Internet was happening, the world of computing became a lot more heterogeneous, and the idea of a mostly uniform world of computing mostly faded away.
> I mean what could be possibly have been wrong with Smalltalk when the IBM's first Java implementations and IDEs were built with Smalltalk
It was handy after IBM's acquisition of OTI.
> Sun bought out the company which developed one the fastest Smalltalk implementations to put their developers into creating Java?
This is inaccurate. Animorphic was bought out for its compiler technology to improve Java; Java was already invented at that point.
> It would be better if there was one vision and one implementation that they all focused on.
Developers are moving on to new things. There is Dart of course (if we are talking about developers of smalltalk), and Kay's VPRI which are attempting to develop the next best thing.
I don't think it was the world of computing becoming more heterogeneous, it was more the desire of the industry majors to push their own backward technologies.
What prevented Smalltalk from being used as the tool for developing web pages instead of Java, other than Sun wanting Java to proliferate by making it free, and IBM and the other Smalltalk developers messing up Smalltalk? And today we have Oracle threatening Java by insisting that copying of APIs is infringing.
What prevented the Smalltalk interface from becoming the browser interface as well as its main development tool? This was the original idea, the original vision the Dynabook vision.
Smalltalk developers should develop it along the lines of Smalltalk/X or Smalltalk/MT and build it all the way up and down as well. Trying to fit it into environments which the industry majors are bent on fragmenting and thwarting in order to buy time put their software, multimedia and storage rental models into place simply won't work.
> Sun bought out the company which developed one the fastest Smalltalk implementations to put their developers into creating Java?
I meant the VM
> Developers are moving on to new things. There is Dart of course (if we are talking about developers of smalltalk), and Kay's VPRI which are attempting to develop the next best thing.
The Dart developer experience will never match the Smalltalk developer experience and VPRI is also building on the crippled Javascript/HTML/DOM foundations.
I can't see why these smart guys don't see that you cannot build on systems which the industry majors are deliberately hobbling in order to buy time to put the rental models into place. They break it and complain that the standards bodies are slowing things down when they and their proxies are the standards bodies.
Smalltalk developers need to knuckle down and get their NPAPI/Pepper/NaCl act together. As that is the only way they can succeed with a browser based act.
This is the source of the frustration that must show through my comments. And then, to steal from seanmcdirmid's comment below:
> But then the Internet was happening, the world of computing became a lot more heterogeneous, and the idea of a mostly uniform world of computing mostly faded away.
I think if there's any chance for the concept that Smalltalk proposed to take hold in the current world, it has to be able to seamlessly interact with existing platforms. At this stage, a fully independant environment does not fulfill that prerequisite.
Some of the Pharo developers want to distance it from the Smalltalk name/association which they believe may be holding it back. There is/has been a long heated mailing list thread on it. http://forum.world.st/a-Pharo-talk-from-a-ruby-conference-td...