[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lojban] Re: CLL and modern Lojban



@timoteios: I would love backward-compatibility too, and I guess it would be good to TRY to do so. Unfortunately, I must slightly disagree with you: this is not 100% possible. Well I'm also a C++ developer, and even in that language the backward-compatibility is not 100%. Remember the old "auto" keyword in "C"? Well, its definition and behavior changed for C++2011! And generally speaking, you cannot just take an old piece of code and expect it would 100% work with the latest tools. I tried it! It may work sometimes, but this is not desirable. Indeed, developers learned a lot about what to NOT DO with old language, and created guidelines to move away from bad practices. And this takes time! This is the reason why "CPP Core Guidelines" exist... So yes, as you say, we should try to have linear history, backward-compatibility, and such. But that's a non-reachable target, IMO.

However, having a committee reviewing the "standard" of the language (as ISO for C++), is just what sounds good to me. Do you think ISO could be the organization to deal with Lojban? :-)

Finally, we should be careful about moving forward to new versions. C++ adopted a 3-years period. Lojban initially adopted 5-years. I feel it is reasonable (even if I tend to use always the latest/beta software for my personal usage!). This leaves time for people to get hands on the new things. And remember there are millions of people knowing C++, and maybe only a few thousands for Lojban... Evolution can't use the same frequency.

Don't you think?

la .sykyndyr.


Le mardi 7 novembre 2017 14:49:25 UTC+1, Timothy Lawrence a écrit :
"You can learn the language described here with assurance that it will not be subject to further fiddling by language-meisters."
- Complete Lojban Language http://lojban.github.io/cll/1/2/

I only use CLL Lojban. I believe that an unambiguous language needs to have a central, singular version to stay unambiguous.

I don't know much about the history of OpenGL but it, from a surface level, seems similar to the history of C++, which I am familiar with.
By and large, new features are added to the C++ language, but every version is backwards-compatible with older versions (and almost entirely compatible with its predecessor, C).
An older compiler might choke on new language features, but a new compiler will always work with old code.

Compiler vendors may introduce their own language features as forks (and they are not considered standard C++), but there is always a standard/"strict" mode that can be enabled. Most good features that compilers introduce have been added to the official version, perhaps changed in a way to integrate them better.

This is what I think Lojban should be like.

I want all new / official Lojban versions to be CLL-compatible.



> "jbo_*" (imagine "jbo_FR", "jbo_EN", "jbo_CA"...),

I do not want there to be "jbo_A", "jbo_B", "jbo_C"... that are all incompatible.

I do want "jbo_1", "jbo_2", "jbo_3"..., where it's a sequential chronology and each is a more improved (but backwards-compatible) version.

A tool could explicitly support "jbo_2" and it would implicitly support "jbo_1". If a tool just says it supports "jbo", it would be presumed to support the latest.



> But this group ("committee"?) should accept the fact that language will evolve, whether they like it or not.

I hope that this means that they accept the language can change, by backwards-compatible design and not evolution via drift.

C++ does not at all "evolve" like a natural language and I don't think Lojban should, either. Because C++ has maintained a centralised standard and continued to update, I think it possible for Lojban to do so.



> What is clear, however, is that people don't want to use CLL Lojban (despite the fact that it is still the most thoroughly documented version).

I do. I think most people want to use the latest version of something. I see the CLL Lojban as the latest version, for the backwards-compatibility reasons mentioned above.

I have seen CLL-incompatible versions advertised on the website and new learners are likely to intuit that it's intrinsically better to choose the "latest version", not knowing that learning the modern versions entails embracing a schism.

> To be frank, I feel a bit betrayed. I feel bad because my hope was that Lojban was more that just an experiment. Someone, please, prove me it is...

This is so true for me, and one of the reasons that I became more quiet (although I am writing an unannounced novel that contains CLL Lojban).

> I just don't want to spend time learning things if they would be thrown away in a few months/years

> sykynder: You mention re-integrating forks back into the core language. How do you change something and then make it the same as it was before?

I believe all official Lojban changes must be backwards-compatible for Lojban to succeed in its goal of being unambiguous (let alone succeed in other ways, such as adoption).



> E.g. learning a revised meaning {lo} is no big deal.

It's not just about learning, it's about effort invested in writing tools, texts, chatbots and parsers (et cetera).
It's about breaking that promise in the CLL that the language will not be subject to further fiddling.

Introducing changes that break Lojban compatibility will
- waste invested time in the older version
- waste invested money in buying the older CLL (etc)
- alienate those who wasted their time/money
- introduce more ambiguity (The Lojban version changes the meaning!!)

(Why did CLL-{le} become modern-{lo} and the modern-{le} get introduced, instead of just introducing the new {lo} to mean what modern-{le} means and keeping CLL-{le} as the default?)

As learning a revised meaning is no big deal, simply redoing these changes to be compatible with CLL Lojban should be no big deal ;)



"Lojban does not yet have nearly the vocabulary it needs to be a fully usable language of the modern world"
- Complete Lojban Language http://lojban.github.io/cll/1/2/

If we can agree to move forward using CLL Lojban and only make backwards-compatible changes, then I can personally say I'll be happy to resume devoting time, energy and money into Lojban. Lojban's usefulness as an unambiguous language is so important to me.

It doesn't have to be "perfect" or "better", it just has to be unambiguous and complete.


I want to help complete Lojban, and any processes of governance surrounding it, to reunite it.

Thanks for reading,


mi'e la timoteios.

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at https://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.