Steve Bennett wrote:
On Tue, Jul 7, 2009 at 5:01 AM, [email protected] wrote:
The reason BASIC was and still enjoys wide popularity is because it's easier to learn.
The example does not make the substantial point because it veers so strongly to the opposite end of the spectrum as to be unrelated to the argument whatsoever. I never suggested that a language should *mimic* English (or a bizarre type of hyper-English).
I welcome however, anyone who wants to actually conduct this argument, on Earth.
The difference between this thread and the parallel one on wikitech-l: that thread quickly focussed on four genuine candidates: Lua, Python, JavaScript and PHP. People identified the basic requirements (security, speed...) and pointed out the pros and cons of each language, in terms of available interpreters, tried and tested experiments with sandboxing each, etc.
Here, we're talking about bringing back BASIC because it's so much more readable. *yawn*
Steve
Can we take this discussion back to wikitech-l now, please, and focus on specific, concrete proposals for syntax reform and/or language replacement?
-- Neil
Hoi, The argument that a language should be readable and easy to learn is REALLY relevant and powerful. A language that is only good for geeks is detrimental to the use of MediaWiki. Our current templates and template syntax are horrible. Wikipedia is as a consequence hardly editable by everyone. Thanks, GerardM
2009/7/8 Neil Harris [email protected]
Steve Bennett wrote:
On Tue, Jul 7, 2009 at 5:01 AM, [email protected] wrote:
The reason BASIC was and still enjoys wide popularity is because it's easier to learn.
The example does not make the substantial point because it veers so strongly to the opposite end of the spectrum as to be unrelated to the
argument
whatsoever. I never suggested that a language should *mimic* English
(or a
bizarre type of hyper-English).
I welcome however, anyone who wants to actually conduct this argument,
on
Earth.
The difference between this thread and the parallel one on wikitech-l: that thread quickly focussed on four genuine candidates: Lua, Python, JavaScript and PHP. People identified the basic requirements (security, speed...) and pointed out the pros and cons of each language, in terms of available interpreters, tried and tested experiments with sandboxing each, etc.
Here, we're talking about bringing back BASIC because it's so much more readable. *yawn*
Steve
Can we take this discussion back to wikitech-l now, please, and focus on specific, concrete proposals for syntax reform and/or language replacement?
-- Neil
Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Gerard Meijssen wrote:
Hoi, The argument that a language should be readable and easy to learn is REALLY relevant and powerful. A language that is only good for geeks is detrimental to the use of MediaWiki. Our current templates and template syntax are horrible. Wikipedia is as a consequence hardly editable by everyone. Thanks, GerardM
I agree. Could you make some concrete suggestion about how you would go about improving it?
-- Neil
Hoi, My instict would be to slash and burn the existing templates. That is not the way to go. So I would introduce a new functionality that has new functionality that is to replace the old crap. Once the new functionality is functional, I would prevent any modifications of the old stuff and replace the old stuff whenever. Priority would be given for replacing the performance hogs. Thanks, GerardM
2009/7/8 Neil Harris [email protected]
Gerard Meijssen wrote:
Hoi, The argument that a language should be readable and easy to learn is
REALLY
relevant and powerful. A language that is only good for geeks is
detrimental
to the use of MediaWiki. Our current templates and template syntax are horrible. Wikipedia is as a consequence hardly editable by everyone. Thanks, GerardM
I agree. Could you make some concrete suggestion about how you would go about improving it?
-- Neil
Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Wed, Jul 8, 2009 at 10:16, Gerard Meijssen[email protected] wrote:
The argument that a language should be readable and easy to learn is REALLY relevant and powerful. A language that is only good for geeks is detrimental to the use of MediaWiki. Our current templates and template syntax are horrible. Wikipedia is as a consequence hardly editable by everyone.
Mortals _use_ the templates, not _create_ them. Geeks create templates for mortals.
Current syntax is indeed horrible, but complete readibility is not the main issue I'd say. Security, speed and flexibility should be, along the ease of implementation.
Peter
Hoi, What is the "ease of implementation" other then being able to use the syntax ? When you say that geeks create templates, you forget that the majority of our projects do not have the geeks to support all these weird and wonderful templates. You forget that there are loads of MediaWiki installations outside the WMF as well. REALLY, the inability of people to do this geek thing is detrimental to the adoption of MediaWiki.
PS I have been a programmer for quite some years, I can submit to the WMF SVN and I will not touch templates with a bargepole when I can help it. I can because typically I refuse. Thanks, GerardM
2009/7/8 Peter Gervai [email protected]
On Wed, Jul 8, 2009 at 10:16, Gerard Meijssen[email protected] wrote:
The argument that a language should be readable and easy to learn is
REALLY
relevant and powerful. A language that is only good for geeks is
detrimental
to the use of MediaWiki. Our current templates and template syntax are horrible. Wikipedia is as a consequence hardly editable by everyone.
Mortals _use_ the templates, not _create_ them. Geeks create templates for mortals.
Current syntax is indeed horrible, but complete readibility is not the main issue I'd say. Security, speed and flexibility should be, along the ease of implementation.
Peter
Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Sun, Jul 12, 2009 at 6:13 PM, Gerard Meijssen[email protected] wrote:
You forget that there are loads of MediaWiki installations outside the WMF as well. REALLY, the inability of people to do this geek thing is detrimental to the adoption of MediaWiki.
You think? What's another wiki that even has user-programmable templates? The MediaWiki templating system is actually a real strength.
Steve
Hoi, Well this strength is not that great when people like myself who has commit right on SVN does not want to touch templates with a barge pole if I can help it. Wikipedia is supposed to be this thing everybody can edit. Thanks, Gerard
2009/7/16 Steve Bennett [email protected]
On Sun, Jul 12, 2009 at 6:13 PM, Gerard Meijssen[email protected] wrote:
You forget that there are loads of MediaWiki installations outside the WMF as well. REALLY, the inability of people to do this geek thing is detrimental to the adoption of MediaWiki.
You think? What's another wiki that even has user-programmable templates? The MediaWiki templating system is actually a real strength.
Steve
Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On Fri, Jul 17, 2009 at 12:06, Gerard Meijssen[email protected] wrote:
Well this strength is not that great when people like myself who has commit right on SVN does not want to touch templates with a barge pole if I can help it. Wikipedia is supposed to be this thing everybody can edit.
I think you misprioritise the whole thing.
Consider it a feature, not a base functionality. Most installations do not use 10% of the possible features available, due to lack of knowledge, time, bravery or else. Writing templates with code is a rare art by my observation, most of the larger MW installations never have used it in the first hand.
You would like to remove TeX (math) input because the language is complex?
And since it'd be an extension I guess, it would imply that you want to forbid(?) creating complex extensions? That's unrealistic. Geeks need this functionality, so ungeeks may or may not care about that, it doesn't really matter. If it's easier to understand than not: that's a plus.
By the way I wouldn't touch PHP with a teen feet pole and rubber gloves, but I'm fine with current template syntax. We're individuals with our own preferences. ;-)
grin
Hoi, Given that there is no alternative to the templates the are imho opinion crucial and core functionality. It is key functionality for everyone who wants to set up a MediaWiki installation that looks good. Tex is effectively a corner case and is not comparable in importance. Thanks, GerardM
2009/7/17 Peter Gervai [email protected]
On Fri, Jul 17, 2009 at 12:06, Gerard Meijssen[email protected] wrote:
Well this strength is not that great when people like myself who has
commit
right on SVN does not want to touch templates with a barge pole if I can help it. Wikipedia is supposed to be this thing everybody can edit.
I think you misprioritise the whole thing.
Consider it a feature, not a base functionality. Most installations do not use 10% of the possible features available, due to lack of knowledge, time, bravery or else. Writing templates with code is a rare art by my observation, most of the larger MW installations never have used it in the first hand.
You would like to remove TeX (math) input because the language is complex?
And since it'd be an extension I guess, it would imply that you want to forbid(?) creating complex extensions? That's unrealistic. Geeks need this functionality, so ungeeks may or may not care about that, it doesn't really matter. If it's easier to understand than not: that's a plus.
By the way I wouldn't touch PHP with a teen feet pole and rubber gloves, but I'm fine with current template syntax. We're individuals with our own preferences. ;-)
grin
Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
2009/7/17 Gerard Meijssen [email protected]:
Given that there is no alternative to the templates the are imho opinion crucial and core functionality. It is key functionality for everyone who wants to set up a MediaWiki installation that looks good. Tex is effectively a corner case and is not comparable in importance.
So your needs are crucial and core functionality, but others' needs are corner cases and not comparable?
I can't think why you're not convincing people.
- d.
Hoi, The needs I am talking about are ALL the MediaWiki installations outside of the WMF. Most, almost all of these do not use Tex. As a consequence you are wrong to consider them "my" needs because it is not about me. Thanks, GerardM
2009/7/17 David Gerard [email protected]
2009/7/17 Gerard Meijssen [email protected]:
Given that there is no alternative to the templates the are imho opinion crucial and core functionality. It is key functionality for everyone who wants to set up a MediaWiki installation that looks good. Tex is
effectively
a corner case and is not comparable in importance.
So your needs are crucial and core functionality, but others' needs are corner cases and not comparable?
I can't think why you're not convincing people.
- d.
Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Hi! Is there any progress on this? Or, that's closed topic and such built-in language won't be implemented? By the way, the lots of MediaWiki installations use TeX, so OCaml is already available as the language for extensions. Imagine #ocaml parser function or an ocaml tag. It's not the most lightweight choise, though. Also, Bryan Tong Minh had a great idea - choosing between an external (compiled) version of Lua (or another scripting language) and "fallback" to slow built-in interpreter, when the first choice is unavailable. Perhaps a limited subset of Lua. Dmitriy
Dmitriy Sintsov wrote:
Hi! Is there any progress on this? Or, that's closed topic and such built-in language won't be implemented? By the way, the lots of MediaWiki installations use TeX, so OCaml is already available as the language for extensions. Imagine #ocaml parser function or an ocaml tag. It's not the most lightweight choise, though. Also, Bryan Tong Minh had a great idea - choosing between an external (compiled) version of Lua (or another scripting language) and "fallback" to slow built-in interpreter, when the first choice is unavailable. Perhaps a limited subset of Lua.
That thread was just talk, I don't think anyone was interested enough to actually do anything. Nobody made any promises.
-- Tim Starling
Tim Starling wrote:
Dmitriy Sintsov wrote:
Hi! Is there any progress on this? Or, that's closed topic and such built-in language won't be implemented? By the way, the lots of MediaWiki installations use TeX, so OCaml is already available as the language for extensions. Imagine #ocaml parser function or an ocaml tag. It's not the most lightweight choise, though. Also, Bryan Tong Minh had a great idea - choosing between an external (compiled) version of Lua (or another scripting language) and "fallback" to slow built-in interpreter, when the first choice is unavailable. Perhaps a limited subset of Lua.
That thread was just talk, I don't think anyone was interested enough to actually do anything. Nobody made any promises.
-- Tim Starling
I did do some experimentation in making JavaScript embeddable. Hit a small snag that I needed help with, but Lua's dev hasn't seamed to have responded to my query.
~Daniel Friesen (Dantman, Nadir-Seen-Fire) [http://daniel.friesen.name]
On Tue, Sep 8, 2009 at 2:02 AM, Dmitriy Sintsov [email protected] wrote:
Is there any progress on this?
No.
Or, that's closed topic and such built-in language won't be implemented?
There was no plausible candidate that would a) be reliably available on shared hosting, b) be acceptably secure, and c) not require us to write our own interpreter in PHP. There is probably no such language that exists. I find it very unlikely that any progress will be made unless at least one of those requirements is relaxed.
By the way, the lots of MediaWiki installations use TeX, so OCaml is already available as the language for extensions.
No, that's not acceptable. You can still use practically all Wikipedia content without getting texvc working. If an embedded scripting language were added, it's a certainty that the large majority of templates on Wikipedia would be gibberish without it. Personally I don't see what's wrong with saying that to fully use Wikipedia content your host needs to allow exec() -- you can get hosting for $3/month that does. But if the user can compile texvc, he can use some sane language like Lua as well, so there's certainly no reason OCaml should be on the table.
Also, Bryan Tong Minh had a great idea - choosing between an external (compiled) version of Lua (or another scripting language) and "fallback" to slow built-in interpreter, when the first choice is unavailable. Perhaps a limited subset of Lua.
It can't be a limited subset if Wikipedia templates are meant to work. It needs to be an exact match. Someone would have to write a Lua interpreter in pure PHP, which is unlikely to happen.
* Aryeh Gregor [email protected] [Tue, 8 Sep 2009 09:34:25 -0400]:
It can't be a limited subset if Wikipedia templates are meant to work. It needs to be an exact match. Someone would have to write a Lua interpreter in pure PHP, which is unlikely to happen.
It's possible to preprocess the source first, to "filter out" the extra language constructions which are not supported by built-in interpreter. Throw an error in such case. Anyway, maybe Victor Vasiliev will present a different language. Dmitriy
On Tue, Sep 8, 2009 at 10:49 AM, Dmitriy Sintsov [email protected] wrote:
It's possible to preprocess the source first, to "filter out" the extra language constructions which are not supported by built-in interpreter. Throw an error in such case.
Then the page will still be broken, so that doesn't help anything.
Anyway, maybe Victor Vasiliev will present a different language.
I believe he is, yes.
* Aryeh Gregor [email protected] [Tue, 8 Sep 2009 10:55:49 -0400]:
Then the page will still be broken, so that doesn't help anything.
It will be broken only if some extra non-supported language constructions are presented in the source text. One syntax check pre-processor for both paths: PHP interpreter and external interpreter. Dmitriy
Dmitriy Sintsov wrote:
Hi! Is there any progress on this? Or, that's closed topic and such built-in language won't be implemented?
I've almost finished writing an interpreter which is implemented using PHP, but can easily ported to C using YACC. I hope it will be ready soon.
--vvv
On Tue, Sep 8, 2009 at 9:45 AM, Victor Vasiliev[email protected] wrote:
Dmitriy Sintsov wrote:
Hi! Is there any progress on this? Or, that's closed topic and such built-in language won't be implemented?
I've almost finished writing an interpreter which is implemented using PHP, but can easily ported to C using YACC. I hope it will be ready soon.
An interpreter for what language?
-Robert Rohde