May be worth to read
especially the call for Squeak tools in Tweak.
This would really help getting Tweak faster. At least it seems to me.

Stef

Begin forwarded message:

> From: "Andreas Raab" <andreas.raab at gmx.de>
> Date: 4 février 2005 13:29:01 GMT+01:00
> To: <tweak at impara.de>
> Subject: [Tweak] A strategic update
> Reply-To: tweak at impara.de
>
> Hi Folks,
>
> Yesterday, I got reminded that it has been a while since I have said 
> anything about the strategic direction(s) Tweak is taking. For those 
> of you following the update stream closely you might have seen some 
> updates hinting on a number of things but since this is sometimes hard 
> to guess here is a bit of a roadmap for the next steps:
>
> Infrastructual issues
> =====================
>
> * Squeak 3.8 and m17n update
> I've spent a good amount of time looking into the issues of updating 
> Tweak to 3.8 in order to get m17n support. I am now ready to get this 
> process rolling (I have already posted a few updates to address a few 
> issues even before they get problematic) and I am expecting to get 
> this done within february (early march at latest).
>
> Note that my current thinking is that the 3.8 migration will only 
> support the 1.1 update stream and NOT the 1.0 update stream. This is 
> because Croquet is still using the 1.0 (3.6-based) version and 
> therefore I will need an update stream for a 3.6-based system and I am 
> planning on the 1.0 stream to be that. In other words, if you want to 
> use a 3.8-based version of Tweak it will *require* you to use the 
> what's currently the 1.1-based updates. If you have a problem with 
> this plan, now would be an excellent time to mention this ;-)
>
> * Using Monticello for code management
> Michael finally convinced me that it might actually be useful for a 
> variety of reasons to give you an option of using MC-based Tweak 
> packages. To this effect, Bert and I have started to break up the 
> Tweak code base in proper packages. This process isn't finished yet 
> but if you are interested check out http://source.impara.de/Tweak.html 
> (see wiki for installation instruction).
>
> We (that is Bert, Michael and myself) fully expect that we will fry 
> the whole set of packages that's up there before everything is over so 
> don't rely on it (also, the packages are currently made for 3.6 - the 
> adjustments for 3.8 have not been made yet) but if you are adventurous 
> enough you might check out the packages to get a feel for the 
> structure of it.
>
> Whether we'll use MC or updates for some of our internal stuff is 
> undecided yet. It might well be that we decide to use updates with 
> eventual new versions pushed into MC for "outside consumption" (this 
> would be very much like "external updates" where we require a grace 
> period for internal testing before we make things available for 
> external use).
>
> Active development
> ==================
> The following areas are the current center of attention in Tweak 
> development.
>
> * Islands: Isolating sets of objects
> Islands is a bit of new technology for Tweak allowing to completely 
> isolate sets of objects from each other. I will send another message 
> to discuss a little more about Islands but for now let it be said that 
> Islands are roughly equivalent to VATs from the E language 
> (www.erights.org) just without imposing a particular message passing 
> paradigm (this is because I want to be able to use different paradigms 
> for interacting islands).
>
> Islands will form the basis for encapsulating projects as well as 
> encapsulating changes to code (e.g., modularity). While I am 
> personally working on the former, Dan is currently looking at the 
> latter.
>
> * Projects and Applications
> Finishing and polishing up the nature of "projects" in Tweak is 
> actually my #1 item on the priority list. Projects will (based on 
> Islands technology) be "strongly isolated" which -for all practical 
> purposes- is equivalent to the level of isolation one usually gets 
> from OS-processes.
>
> Projects will be able to run concurrently (and preemptively), they 
> will be capable of being checkpointed and restored, they can be 
> "killed" (with automatic cleanup of resources used) etc. In other 
> words, using Islands projects will have all of the properties that one 
> usually expects from "applications" and they will function precisely 
> that way.
>
> Some time further down the road (and there is no deadline for it yet) 
> projects will also be able to carry its own code along with it and be 
> able of using this code without interference with any other projects 
> (yes, I know it sounds crazy but Dan and I talked about this and we 
> both believe this is perfectly possible).
>
> * Cairo support and a sane graphics interface
> Some of you might know that Ned is working on a Cairo 
> (www.cairographics.org) integration for Tweak. As soon as we have 
> versions of Cairo which can be compiled on the major platforms I am 
> planning to do some major reformulations of the current Canvas 
> interface into a stock-of-the-mill 2D graphics interface. Why? To 
> support printing. When this reformulation is complete, all of Tweak 
> WILL print out of the box to Cairo-supported backends, and possibly to 
> others as well.
>
> * End-user tools
> We're making good progress towards implementing the missing bits and 
> pieces for end-user tool support from Squeak. Once we're complete I 
> think we can start to give some other apps (say OpenOffice Draw) a run 
> for its money ;-) Ted and Scott are both heavily involved in these 
> areas and making great progress.
>
> * Tile-Scripting
> The same here. We're in the process of finishing up the remaining bits 
> and pieces and make it so that we get (at least) back on par with what 
> used to be in Squeak. Currently, Bert and Ted are driving forces here.
>
> * Scripting syntax
> Another significant area of development will be scripting syntax - 
> this is currently on #2 of my personal priority list right after 
> finishing projects so expect something in this area in the coming 
> months.
>
> Cool stuff
> ==========
> While not exactly critical for Tweak itself, there are plenty of cool 
> things going on right now, here are a few:
>
> * Areitha Ffenestri support
> With the above mentioned project architecture I am fully expecting 
> that we will be trivially capable of supporting multiple windows (aka: 
> Areitha Ffenestri) in Tweak. Multiple windows will effectively be just 
> multiple projects. Bernd, Michael and of course Tim, are the driving 
> forces in this area.
>
> * Filters, Tasks, Interactors
> We have just started to explore a new architecture in Croquet which we 
> call "filters, tasks, and interactors" and which is basically the 
> definition of an application as a contextualized set of tasks 
> performed through a (2D or 3D) filtering view of the world (there is a 
> paper about it in this years C5 but it isn't online yet; I hope we'll 
> have it online at the Croquet site before too long). I find this an 
> amazingly refreshing view on any kind of authoring environment and I 
> am planning to change most of Tweak's authoring environment to support 
> the same metaphor. This will be an ongoing project over the next 
> months since we're just starting to understand this architecture in 
> Croquet.
>
> * Widget Design overhaul:
> I have recently seen some interest to do a bit of widget design 
> overhaul to make the current widget set "more modern". Since I liked 
> what I saw in a few early screenshots I expect that we might have (one 
> or more) really nice looks perhaps around mid-year.
>
> On the back-burner
> ==================
> Here are a few projects which people might expect to see active 
> development in but which are unlikely to see much attention unless you 
> ask for professional help (which can, but needn't be Impara).
>
> * Smalltalk development tools:
> There is current NO plan for "completion" of the Smalltalk development 
> tools. We fix a few things here and there when we have an itch to 
> scratch but there is currently no plan to complete browsers, 
> debuggers, change sorters, monticello etc. etc. etc.
>
> If you have a need for those tools in Tweak you have the following 
> options:
> a) Get professional help. Porting tools such as one of the many 
> browsers, or the MC interfaces isn't exactly hard (it just takes time) 
> and I am sure there are people out there which will do it for a 
> reasonable price.
> b) Set up a team of interested people and port the stuff you need for 
> yourself. I have said this in the past: The first person who gives me 
> a fully working browser in Tweak is "in". No questions asked. It just 
> needs to work. If you like Omni, Whisker, Star or any other browser 
> and want it to be the default in Tweak, just be the first at porting.
>
> * Native Widgets:
> We have played with native widgets in the past and while much of this 
> was pretty successful it is simply a daunting task to deal with. If 
> you need native widgets and if you need them in Tweak I would 
> currently recommend looking at wxSqueak and the potential of using 
> wxWidgets as costumes for Tweak players. Given the experience we had 
> so far this seems like the best option (and quite doable).
>
> Cheers,
>  - Andreas
>
> _______________________________________________
> Tweak mailing list
> Tweak at impara.de
> http://impara.de/mailman/listinfo/tweak
>