Discontinuing projects

[TR NS]

If anyone wants the source to any of these three projects please fork them ASAP.

* [YAML DOM](https://github.com/trans/yamldom.cr)
* [Neapolitian](https://github.com/trans/neapolitan.cr)
* [Ruby.cr](https://github.com/trans/ruby.cr)
* [Brite](https://github.com/trans/brite.cr)

I am discontinuing the projects and will be deleting the repos soon.

Note that YAML DOM is the most mature of the three and is basically "almost done", but there are some difficulties with finishing it off b/c of #each and the Crystal type system.

The next three work in so far as I got with them, but are very nascent.

Brite uses Neapolitan, btw. And I had plans of making Brite more robust/flexible by using YAML as an intermediate representation for generation of a site (which is why I worked on a YAML DOM), but its become clear that's not going to be possible any time soon (if ever).

[TR NS]

Oh, p.s. don't take all of the READMEs for face value -- they were still works in progress and don't necessarily match the code or evolving designs 100%.

[Chris Hobbs]

Why would you delete the repos as opposed to just leaving them with a discontinued notice?

--
You received this message because you are subscribed to the Google Groups "Crystal" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crystal-lang...@googlegroups.com.
To post to this group, send email to crysta...@googlegroups.com.
Visit this group at https://groups.google.com/group/crystal-lang.
To view this discussion on the web visit https://groups.google.com/d/msgid/crystal-lang/e2a9899e-bbb1-43a3-8706-d5c35683df00%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ryan Gonzalez]

:( I kind of get the impression that you didn't like Crystal's type system too much...

--
You received this message because you are subscribed to the Google Groups "Crystal" group.

To unsubscribe from this group and stop receiving emails from it, send an email to crystal-lang+unsubscribe@googlegroups.com.

To post to this group, send email to crysta...@googlegroups.com.
Visit this group at https://groups.google.com/group/crystal-lang.
To view this discussion on the web visit https://groups.google.com/d/msgid/crystal-lang/e2a9899e-bbb1-43a3-8706-d5c35683df00%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Ryan (ライアン)

Yoko Shimomura > ryo (supercell/EGOIST) > Hiroyuki Sawano >> everyone else

http://kirbyfan64.github.io/

 

[Tim Uckun]

You clearly put in a lot of work on these. I suggest you keep them up there in case somebody in the future wants to use them or learn from them. 

On Tue, Dec 13, 2016 at 7:51 AM, Chris Hobbs <ch...@rx14.co.uk> wrote:

Why would you delete the repos as opposed to just leaving them with a discontinued notice?

On 12/12/16 18:24, TR NS wrote:

If anyone wants the source to any of these three projects please fork them ASAP.

* [YAML DOM](https://github.com/trans/yamldom.cr)
* [Neapolitian](https://github.com/trans/neapolitan.cr)
* [Ruby.cr](https://github.com/trans/ruby.cr)
* [Brite](https://github.com/trans/brite.cr)

I am discontinuing the projects and will be deleting the repos soon.

Note that YAML DOM is the most mature of the three and is basically "almost done", but there are some difficulties with finishing it off b/c of #each and the Crystal type system.

The next three work in so far as I got with them, but are very nascent.

Brite uses Neapolitan, btw. And I had plans of making Brite more robust/flexible by using YAML as an intermediate representation for generation of a site (which is why I worked on a YAML DOM), but its become clear that's not going to be possible any time soon (if ever).

--
You received this message because you are subscribed to the Google Groups "Crystal" group.

To unsubscribe from this group and stop receiving emails from it, send an email to crystal-lang+unsubscribe@googlegroups.com.

To post to this group, send email to crysta...@googlegroups.com.
Visit this group at https://groups.google.com/group/crystal-lang.
To view this discussion on the web visit https://groups.google.com/d/msgid/crystal-lang/e2a9899e-bbb1-43a3-8706-d5c35683df00%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Crystal" group.

To unsubscribe from this group and stop receiving emails from it, send an email to crystal-lang+unsubscribe@googlegroups.com.

To post to this group, send email to crysta...@googlegroups.com.
Visit this group at https://groups.google.com/group/crystal-lang.

To view this discussion on the web visit https://groups.google.com/d/msgid/crystal-lang/fcaaf10a-5a53-a4f6-db8e-da995451a558%40rx14.co.uk.

[TR NS]

Sorry about the late reply. At first I wasn't sure I would reply, but I was thinking about it today, and I thought maybe others deserve to at least have a basic idea as to why.

On Monday, December 12, 2016 at 3:30:18 PM UTC-5, Ryan Gonzalez wrote:

:( I kind of get the impression that you didn't like Crystal's type system too much...

I would say that is partly true, but not the primary thing. Mostly the type system is great. But there were a couple of times where it would have been nice if it were possible to instruct the compiler to handle some dynamicism. That was particularly true with parsing YAML. I needed to create an API that could be easily extended, and the way Crystal is I basically had one interface option: a pull parser. If you think a method that can return different types is prone to error because it isn't the "static way", consider how bug prone a pluggable pull parser would be. I managed to work my way around the type system enough to provide another option -- a much nicer API something like you would see in Ruby. To do it, I simply made Scalar, Sequence and Mapping all have the same interface. This made the compiler happy. But it turns out it isn't possible with Array and Hash #each methods -- the two are completely non-polymorphic. They really don't need to be -- an array index is analogous to a hash key. But when I suggested it, the idea was rejected out of hand. So my API would never be 100% because of this, which basically made the whole thing moot. I could have just settled with a less intuitive, less easy-to-use pull parser API, I may have, but then, and I think this was the final straw: https://github.com/crystal-lang/crystal/pull/3608. I did the work to allow the YAML library to emit tags (and more or less there isn't really any other way to do it) only to be told "don't have time" and they didn't see any reason to fully support YAML spec anyway.

[Leeland Morgan]

Hey come on, just bear with us here. If you want to make Crystal better - don't ragequit please!

четверг, 12 января 2017 г., 20:04:55 UTC+3 пользователь TR NS написал:

[Tim Uckun]

On Thu, Feb 2, 2017 at 4:22 AM, Leeland Morgan <idch...@gmail.com> wrote:

Hey come on, just bear with us here. If you want to make Crystal better - don't ragequit please!

To stick up for him..

He did try very hard and did  a lot of work. He had discussions about his problems here and in the end the language was not going to be changed and what he wanted to do was not possible in Crystal.