REBOL great but flawed => Fw: [Flashcoders] Request for Michael - Fo
[1/6] from: petr:krenzelok:trz:cz at: 26-Jan-2002 21:08
Jason Cunliffe wrote:
>Hi >..some heavy discussion recently on the excellent flashcoders list re:
<<quoted lines omitted: 10>>>>1. Lack of 400 Million Plus users >>
let's hope Morpheus deal will help here and Rebol starts to be widely recognized, although - let's get to the poin no 2. ...
>>2. Lack of strong and growing developer base >>
Acceptance of Rebol can be imo collection of following point: - for some ppl - rebol is not open-sourced, maybe some part of developer community fears of RT disappear or buyout ....(look at Omnimark which I mentioned - now seems to be bought by some company and suddenly closed product ...) - library and shell components are being put into commercial versions, not free ones. Security issues? Maybe. But on the other hand - ppl has to pay for features found for free with other languages (Python, Perl ...). Free Rebol versions are limited to only TCP kind of communication ...
>>3. Large security problem - Rebol applications can access the local >>machine. >>
well, I am not security expert. Rebol lives in "sandbox". If it wants to access externall directory, it requires user to confirm the action (I can show you java script code which can do some very nasty things to your computer, e.g. run "format C:" in IE 5.5 + , by just visiting some website), so judge for yourself ...
>>It is a great application, but there is a high barrier to entry to the >>client side without distribution. For Flash, the market for the new
<<quoted lines omitted: 3>>>>a death nail. >>
well, I agree here. RT claims that their touch with Windows registry for e.g. is minimal. That may be true, but I found several times difficulty even when using -i switch. I don't remember the case exactly, but it had something to do with running script from non-rebol directory. Installation dialog popped-up ... I work as e-business manager for our company and most of companies presenting us their products have one strong selling-point - you don't need to install anything, it just uses user's browser. Managers tend to like the idea ;-) That's why I don't consider Rebol browser plug-in being a plug-in at all, unless it knows how to draw into browser's window area. I imagined something as Oldes showed us with his Rebol Flash dialect ... The situation with Rebol could look being bad from what I said, while I think it is just opposite: RT surely knows advantages of some points I made. However - they choosed different strategy. They will concentrate themselves on the collaborative market. That seems to be the area, where compainies will spend their money for some years now. You can bet that Carl & co. wants the best for Rebol, so do I and many of us here. - We just differ in priorities :-) Many of us would like to see Rebol in multimedia market, browser plugins, simply attracting or making ppl dependant upon Rebol thru mass acceptance :-) As RT now concentrates on business users, let's hope new Rebol programmers will come from this area. If not - we are in danger, as currently there is very few Rebol developers/consultans/books available, and IOS is not going to help us to enlarge Rebol's user base ... -pekr-
[2/6] from: greggirwin:mindspring at: 26-Jan-2002 14:28
Hi Jason, Here are some thoughts:
> Rebol is great but it has several fatal flaws. > > 1. Lack of 400 Million Plus users
If this is a fatal flaw, REBOL is in good company. :)
> 2. Lack of strong and growing developer base
Hmmm. I suppose we could be insulted by that, but they are probably only considering the view that strength lies in numbers, which we all know is not the case. I could be wrong, but I think the developer base *is* growing.
> 3. Large security problem - Rebol applications can access the local
machine. I have quite a different view than a lot of people here, and elsewhere. I stayed away from Java for, among other reasons, the *lack* of access to the local machine. Maybe it's because of my mindset, experience, and target applications, but when I played with it originally, that frustrated the heck out of me. I like REBOL's sandbox approach and I think the design choices RT has made work very well. I imagine they have a slightly different design target than some other tools but that's their call. If access to the local machine is a "fatal flaw", REBOL is, again, in good company. Just because you don't agree with a design decision doesn't mean it's a problem.
> It is a great application, but there is a high barrier to entry to the > client side without distribution. For Flash, the market for the new > version market lags 1 year behind the release date, it is a tough sell > to require the download of a new flash player. To tell a consulting > customer that an end user has to download and install something is often > a death nail.
I can tell you why *I* don't want to install a new version of Flash (or any version for that matter). I'm on a dial-up connection and the main use of Flash seems to be animated splash screens that, themselves, take time to download and provide no value to me as a user. I'm used to building applications that provide a fairly rich user experience. The effort required, lack of responsiveness and robustness, and interaction limitations of most "no install required" approaches I've investigated (and experienced as a user) has kept me very clear of entering that camp. I don't want to work harder in order to provide a less satisfying user experience just so I can say "no install required". That's my personal preference. Just my opinionated opinions. :) --Gregg
[3/6] from: chris:langreiter at: 27-Jan-2002 3:01
Prolog: I don't use REBOL because it is as good as Python or Perl, but because from my experience it is vastly superior - it simply fits my mind better. The parse dialect is a godsend and I really start to appreciate the power of /View. However ...
> - library and shell components are being put into commercial versions, > not free ones. Security issues? Maybe. But on the other hand - ppl has > to pay for features found for free with other languages (Python, Perl > ...). Free Rebol versions are limited to only TCP kind of communication
... I'd really, really, really put those into the "free" versions as well (which, nota bene, aren't free for commercial use either). Just look how the Python world is blooming with modules incorporating everything from OpenGL (ever seen VisualPython - amazing, amazing) to K/Kdb (just announced: http://pin.to/pyk). And it makes sense - bootstrapping is often easier if you can build on (lots of!) other people's work - and be it for the meantime only. Let's stand on shoulder of even more giants. Just my 2 ?Cents. -- Chris
[4/6] from: jason::cunliffe::verizon::net at: 27-Jan-2002 11:07
Thanks to all who answered .. I'll try to summarize and include my own arguments in a later post.. Chris wrote:
> I'd really, really, really put those into the "free" versions as well > (which, nota bene, aren't free for commercial use either). Just look how
> Python world is blooming with modules incorporating everything from OpenGL > (ever seen VisualPython - amazing, amazing) to K/Kdb (just announced: > http://pin.to/pyk). And it makes sense - bootstrapping is often easier if > you can build on (lots of!) other people's work - and be it for the
> only. Let's stand on shoulder of even more giants.
Yes! ...Python is very rich these days [http://www.vex.net/parnassus] Do you mean VPython [was Visual Python] ? http://www.vpython.org/ I cannot find Python interface to K/Kdb: please tell me more.. Do you mean K as in 'KDE linux window manager' based on Trolltech's Qt: http://www.trolltech.com or K/Kdb as in 'Kx Systems' lightening ***fast** data analysis tools http://kx.com/ ./Jason
[5/6] from: chris:langreiter at: 27-Jan-2002 22:45
> I cannot find Python interface to K/Kdb: please tell me more..
> K/Kdb as in 'Kx Systems' lightening ***fast** data analysis tools > http://kx.com/
Here's the link: http://pin.to/pyk It was in the original mail as well ;-) Bye, -- Chris
[6/6] from: jason:cunliffe:verizon at: 27-Jan-2002 18:10
> > I cannot find Python interface to K/Kdb: please tell me more.. > [...] > > K/Kdb as in 'Kx Systems' lightening ***fast** data analysis tools > > http://kx.com/ > > Here's the link: http://pin.to/pyk > > It was in the original mail as well ;-)
yes, but the url still does not work. I found this: http://www.thekompany.com/projects/pykde/ That's the Python bindings for KDE/Qt What I am wondering about http://kx.com/ ? thanks ./Jason
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted