Does RPerl have any users? This is a one-man project and I am a little surprised to see that the Perl Foundation has already put funds towards it. If the author were working with the Perl 5 maintainers then that would be one thing, but he seems to be on his own.
There is a language that I'd like to see go faster, and that is Perl 6. Please spend money on that.
Well Perl 5 was originally a one man project!
I was impressed with Will's talk at FOSDEM 2016 and hope it attracts more support.
I think its worth supporting perl 5, perl 6 and rperl and not putting all our eggs in one basket.
Do you have written permission from the copyright holder(s) of Learning Perl to do this? What are the terms of the license/waiver?
"Does RPerl have any users?" - Me!
That's why I need the documentation to be 100% complete.
I'm sure when docs finish and are 100% fully edited, more users will come to RPerl.
I have friends that showed interest in RPerl, but due to the lack of detail documentation, they are hesitant.
Please support Will, because his project deserves it.
Definitely needed, RPerl have future
We need cool projects like this to keep perl 5 alive and kicking, and high quality documentation seems to attract more users.
I vote for this project.
Investigating performance opportunities requires trying different approaches. RPerl is one approach, and Will seemed to eagerly follow his vision for the last years. I'm +1.
RPerl can't expect a proper shot at wider adoption without good documentation, so this task merits support.
I've voted for RPerl before, and also put some money in it, and I wish this grant will be rewarded. RPerl can be so very useful for the evolution of Perl.
Voting YES! Fund it!
RPerl has potential to be a faster , follow up grants may bring Perl5 and RPerl closer together.
Having a Perl that finally allows typing (yes, that means more robust code in larger projects) and brings speed improvements will definitly help.
So if Perl5 wants to stay relevant, the experiments of RPerl and CPerl should be considered a good source for potential to integrate into the main Perl5 branch (or rethink how Perl5 may evolve).
So I vote YES.
One of my successful networking contacts is waiting and waiting for RPerl to be more fully developed, incorporating massive speed-up of large scale data manipulation for his pending projects. He obviously needs to understand RPerl to make use of it in developing a number of smart phone apps that cannot otherwise work. The user manuals for RPerl are crucial. Please support this valuable project!
I Vote YES. Let's get this done!
This is a worthwhile project, and an outgrowth of the good "perl ii" explorations. RPerl has staying power, I'm for funding it.
Yes
+1
I vote for this project too.
I'd say fund. There is not much else going on in Perl5 right now and this is one of the more interesting projects.
I do vote for not only this proposal, but also everything helping me to use the immortal Perl language. And I am looking forward to see the outcome.
I support this grant, rperl is a good thing for the perl community.
RPerl adds another perspective about perl performance and I would like to see this project evolving.
I vote Yes, please fund it!
I vote yes. I hope it will help to the RPerl team to improve the tool.
Performance matters. I've donated RPerl last time on kickstarter, and this project should go further.
Documentation matters. My friends wanted to try RPerl but stopped due to lack of documents.
Voting YES. Fund it!
I vote fund
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
To Whom It May Concern,
I'm quite interested in learning more about RPerl. It sounds like a
neat project and I know that Will is quite dedicated. I've never seen
such an unending run of work on a source code repo (560 days
uninterrupted to date). 60 days worth of work in exchange for $1600
sounds like a steal to me and I advise The Perl Foundation to fund
this project.
Sincerely,
C.J. Collier
cjcollier@cpan.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=t2bS
-----END PGP SIGNATURE-----
I've written a lot of Perl. I would love to keep using my Perl skills for large-scale data processing and scientific computing, and that really means RPerl. R is fast, but it's a language design nightmare. The Python scientific stack is pretty sweet, but using it means digging through stacks of reference documentation to find things about Python that I already know about Perl. Perl is magnificent, but there are speed issues. RPerl seems to be magnificent AND fast ... if only it had complete docs I could rely on. Please fund.
I like RPerl's concept, and the documentation could indeed use some extra work. I vote +1
Yes, I would love to see this happen! It's much vital for Perl 5. I would use it.
I think the man deserves support!
Please fund it!
My vote is YES..
Wil is really passionate about Rperl and after seeing his demos and long term dedication to the project I strongly feel he deserves the Perl Foundations support. 'Classic' Perl needs to maintain long term backward compatibility which makes the language increasingly less viable for new projects the require features like crazy high performance and concurrency. Rperl is a massive opportunity to let people that like Perl and are building new things to get some of these new features, all while not tossing out all hopes of compat with existing code.
I've personally supported the Rperl kickstarter and think the money he's asking for to improve docs and get more people running with it is not a lot for value he's adding.
Compiling Perl into executable binary is definitely useful
I think with RPerl, we could have the faster program without much gain. So I hope documentation of this project is needed.
Voting Yes. :-)
Tutorial material for RPerl would be welcome. Assuming no copyright issues with use of "Learning Perl" this has my vote.
Definitely needed.
I vote Yes, please fund it!
as long as it is done with an eye towards perl 6 compatibility every project that could help gain the interest of the masses (and we never know which projects those turn out to be) is something we should all e for.
+1 for this project!
I vote yes.
RPerl is crucial. Say yest to it and please have it funded. Thanks.
I vote yes.
Voting yes. :^)
Without docs nobody will try to use software, especially rperl. I support this grant.
I vote yes, I would like to see RPerl evolving
Will,
The Grants Committee members have been discussing this. I thought it'd be helpful if you could answer these questions:
1. How is the current user base of RPerl? Any user feedback from users?
2. What's the end goal of the RPerl project? Is it going to co-exist in parallel with Perl 5? Or was there formal/informal discussion with the p5p people to port some work? (I understand RPerl runs in a very different way from Perl 5) In your view, where is RPerl (going to be) located in the Perl ecosystem?
3. Related to the previous question, where is the latest development roadmap? What's next after the documentation is complete? How far are you from the whole project completion?
Thank you,
Yes for funding! RPerl has improved perl and can sustain against php/c++ in terms of performance!!!
Hi,
+1
I vote for this project as I have been using RPerl since the beta version.
Thanks,
My vote is YES
OK, why not. Seems fair.
RPerl is one approach to making Perl grow as a community and good quality documentation is a major attractor. I vote yes.
The grant application starts:
The number one request and obvious need at this time is still quality RPerl user documentation, to help new RPerl users learn how to write fast software.
There's a line in the "Details", Exercises in Learning Perl which are not supported by RPerl are omitted. I don't have a copy of Learning Perl, hence I don't know how many exercises it has in total. Hence another (small) question:
How many Learning Perl exercises were omitted because they are not supported by RPerl? (to give a feel for how likely folks are to be unable to use RPerl because it's incomplete)
I vote YES, RPerl has the potential to boost up Perl5 in speed and popularity.
I vote for this project
I would like to see the project receives all the funding as needed, I have seen exceptional commitment by the author so far and I am very excited about future efforts.
I am a user of RPerl and I have contributed via the kickstarter project page some months ago.
Greetings TPF Grant Committee Members & Fellow RPerl Enthusiasts! I am humbled and inspired by everyone's huge show of support. To answer your questions so far:
1. How is the current user base of RPerl? Any user feedback from users?
The primary feedback I have received are (unsurprisingly) requests for user documentation, which is the subject of this grant proposal. As you can see from the multitude of comments on this grant proposal blog post, there is a clear mandate from the public Perl user base. Sue Spence implied a lack of community interest in RPerl by (perhaps sarcastically) posing the question "Does RPerl have any users?", the dozens of responses to which have obviously proven her wrong. I am not aware of any mechanism to track every RPerl user, I could estimate anywhere from dozens to hundreds of users on darkPAN. Nobody should question the community's interest in RPerl after witnessing such a positive outpouring of support in the form of supportive blog comments.
2. What's the end goal of the RPerl project? Is it going to co-exist in parallel with Perl 5? Or was there formal/informal discussion with the p5p people to port some work? (I understand RPerl runs in a very different way from Perl 5) In your view, where is RPerl (going to be) located in the Perl ecosystem?
This is an open-ended and complex question, the full answer to which can easily take up a few hours and is the subject of my new presentation "RPerl, Perl 11, and The Future Of Perl Performance" as debuted a few weeks ago at FOSDEM in Brussels. The short answer to "the end goal of RPerl" could be "to massively speed up Perl 5 and eventually Perl 6"; the long answer is similar, with many additional layers of technical variations and detail. One of several possible outcomes is to re-write both Perl 5 and Perl 6 from scratch using RPerl, which would allow us to maintain full "bug-for-bug" backward compatibility with both XS and CPAN, while simultaneously achieving 2 major goals: unbeatable performance, and reunification of P5 w/ P6. Advanced topics of discussion in this area include RPython, PyPy, PONIE, cPerl, P2, Perl 11, etc.
RPerl is a low-magic subset of Perl 5, and runs using Perl 5, so yes I think it is safe to say that RPerl "co-exists in parallel with Perl 5", although some might argue wording and say RPerl exists "within the realm of Perl 5". When RPerl reaches an appropriate level of maturity, we will approach p5p to get RPerl included in the Perl 5 core. As RPerl stands now, it is an independent development project and should be funded alongside the existing Perl 5 and Perl 6 "rolling grants". RPerl is, by all meaningful measurements, Perl core system software, and should be treated as such.
3. Related to the previous question, where is the latest development roadmap? What's next after the documentation is complete? How far are you from the whole project completion?
The current RPerl development roadmap can be found here:
https://trello.com/b/hQCPeg1d/austin-pm-rperl-development-roadmap
After basic Learning RPerl documentation is complete, there are several groups of tasks to be tackled (ignoring for now the much larger jobs of rewriting P5 or P6):
- RPerl v1.x System Code (remaining C++ code generation & tests, various misc items)
- RPerl Benchmark Apps (Alioth, sorting, fractals, etc)
- RPerl User Apps (math, science, big data, web services, databases, etc)
- RPerl v2.x System Code (regular expressions, parallelism)
- RPerl v3.x System Code (medium-magic support)
There are multiple possible answers to when "whole project completion" will occur; one possible short-term answer might be "when RPerl v3.x is done with medium-magic grammar support" or perhaps even "when RPerl v4.x is done with high-magic support". Another likely answer might be "when Perl 5 and Perl 6 are both rewritten from scratch, are reunified, and run as fast as C++". Yet another valid answer could be that RPerl will be completed "when Perl 5 and Perl 6 are complete"; obviously, both P5 and P6 are ongoing, open-ended projects with no completion dates planned or intended. When will RPerl be done? You may need to first ask the question: when will Perl be done? I don't know the answer to that one.
Also, Mr. Perrett asked: "Do you have written permission from the copyright holder(s) of Learning Perl to do this? What are the terms of the license/waiver?"
The book Learning RPerl does not include any copyrighted content or intellectual property from the book Learning Perl, thus no license or waiver is needed. Instead, the book Learning RPerl includes reimplemented-from-scratch low-magic solutions to the exercises in Learning Perl; we do not include any of the original high-magic exercise solutions or even the original exercise questions from the book Learning Perl. We are being very considerate of other peoples' IP and are being very careful not to include other peoples' copyrighted material in Learning RPerl. I don't personally know Tom Phoenix, but I am friends with Randal L. Schwartz and brian d foy, so they are aware of our ongoing plans with the TPF grants for Learning RPerl.
...
Do you have any followup questions for me?
:-)
Perling,
~ Will the Chill
I would like to see the Perl Foundation support chillbert in his rperl efforts. There are few people on this planet more knowledgeable and enthusiastic about all things Perl. Financial support for documentation will lead to more rperl exposure, use and contributions by others. Support the guy's work.
+1 from me also.
I'm really interested in having a faster Perl AND knowing how to make use of it
Voting YES.
Perl is the best thing in programming life.
Rperl could be the same in Perl life.
I totally support RPerl!!
My vote is a big YES!!!!
I am a professional computer programmer with a long-term interest in RPerl. I believe Sue Spence is wrong, it is important to the future of Perl that we use TPF resources to support RPerl development alongside Perl 5 and Perl 6.
Right now, the most needed RPerl work item is user documentation, the book Learning RPerl will be a very valuable base of knowledge for the Perl community. I STRONGLY VOTE YES ON THIS AND ALL FUTURE RPERL GRANT PROPOSALS.
I think having a higher-performance restricted subset that reuses Perl skills is a great idea. TPF already welcomes the idea of two sister languages. Perhaps once RPerl is fully documented someone could make a Perl6 implementation that ties to it, too. That language in particular has been welcoming of multiple implementations.
Will -
In your comment above, you say:
Another likely answer might be "when Perl 5 and Perl 6 are both rewritten from scratch, are reunified, and run as fast as C++".
Have you talked to any of the core Perl 6 developers about this goal? This doesn't jibe with the direction of the Perl 6 project, nor the effort gone into an optimizable runtime, nor Jonathan W.'s efforts upcoming this year to make those optimizations a reality. As far as I know, no one in the Perl 6 community is thinking about a rewrite. (A different implementation is fair game to be called Perl 6, of course, but that's not a rewrite)
The two languages, while they have a lot in common, are different for a reason, and unifying things at the language level at this point is not just a backend/VM issue, but a specification issue of the languages themselves, and at this point both languages have backward compatibility issues which would make unification at that level... difficult.
Mr. Coleda,
Thank you for your interest in RPerl!
You are 100% correct in your assessment that a "different implementation is fair game to be called Perl 6"; when I say "re-write" that is precisely what I mean, creating new implementations of the Perl 5 and Perl 6 specifications, by using RPerl for both instead of using C for P5 and NQP for P6. We are literally saying the exact same thing.
Again, you are 100% correct in stating both Perl 5 and Perl 6 have "issues which would make unification at that level difficult"; it's taken decades to get here, one might imagine decades of work ahead of us to sort it all out.
None of the RPerl plans are meant to conflict with the Perl 6 team's efforts in any way whatsoever. RPerl has a lot to deal with in terms of Perl 5, it will be quite some time before we are ready to address Perl 6!
So basically I agree with everything you've written.
Thanks,
~ Will
Hi Mr. Clark,
Thank you for your interest in RPerl! My early e-mails with you in 2012, which I believe you submitted to TPF as grant hours, were instrumental in my understanding of the many issues involved in compiling Perl. You were the very first person I spoke at the beginning of my quest, just before I connected with Ingy, Reini, and others in the Perl community.
To answer your questions...
Where *are* RPerl's users? It's really not obvious.
There are approximately 60 favorable comments on this grant proposal. In contrast, there are only 14 comments on the Perl 6 Performance grant proposal link, which you provided. If your question is "Where *are* RPerl's users?" then we should also ask "Where *are* Perl 6's users?" The answer is obvious, the RPerl users are right here, in this grant proposal blog thread!
Of the many comments on this grant, *none* say that they are currently using RPerl.
SORRY! This one is factually incorrect. Some examples:
Stefanos | February 15, 2016 8:47 AM
"Does RPerl have any users?" - Me!
prajith | February 17, 2016 4:42 AM
Hi, +1, I vote for this project as I have been using RPerl since the beta version.
Isaac Close | February 17, 2016 11:15 AM
I would like to see the project receives all the funding as needed, I have seen exceptional commitment by the author so far and I am very excited about future efforts.
I am a user of RPerl and I have contributed via the kickstarter project page some months ago.
RPerl's documentation points to the IRC channel #perl11 (which is also used for other projects such as cperl), but the logs are mostly bots reporting build results.
The publicly-logged IRC channel goes through waves of activity. Most RPerl support requests are received directly through e-mail or Facebook messenger. It is not uncommon for me to receive multiple RPerl support requests in a single day, or have a few days with no support requests. This seems common for any early-stage software development project. RPerl has been in active development for about 3 years. Perl 6 has had almost 15 years. I think RPerl is doing just fine comparatively!
It seems that folks only comment on the grant applications. Why do people only comment on the application, but never give feedback on the results?
Why is this strange? When there are anti-RPerl people trying to convince TPF to deny our grant requests, it only makes sense for RPerl's numerous supporters to make their collective voices heard. This is how the community votes. In fact, that is the explicitly stated purpose of the grant applications, according to TPF!
"Before the Committee members vote, we would like to solicit feedback from the Perl community on the proposal."
On the other hand, no such language requesting any feedback was included in any of the grant progress reports. In fact, the exact opposite happens to be true in this case, due to (apparently ongoing) technical issues with the Movable Type blog software:
"Owing to issues with TPF blog, this report could not be published in a timely fashion. The Grants Committee elected to vote on completion without public comments. The committee accepted the work as successfully completed."
Just for fun, let's see how many comments are posted on the other recent TPF grant progress reports!
Perl 6 Release Goals: Final Grant Report ZERO COMMENTS
Grant Report: Test::Simple/Stream ZERO COMMENTS
Maintaining Perl 5: Grant Report for November ZERO COMMENTS
Migrating blogs.perl.org: Second Progress Report ZERO COMMENTS
Maintaining Perl 5: Grant Report for October ZERO COMMENTS
Grant Report: Test::Simple/Stream Stabilization ZERO COMMENTS
Maintaining Perl 5: Grant Report for September ZERO COMMENTS
So yeah, zero comments on grant progress reports is the norm. This argument against RPerl also seems to be totally invalid.
Almost no-one is blogging about RPerl.
SORRY! This one is also factually incorrect. Some examples:
Blogs.Perl.org Search for "RPerl"
FOSDEM 2016 by Mark Keating
FOSDEM 2016 by Steve Mynott
Fixed 5.22 problems during my compiler port by Reini Urban
What's the perl5's future? by xiaoyafeng
On Dave Mitchell calling B and B::C a failed experiment by Reini Urban
How many Learning Perl exercises were omitted because they are not supported by RPerl?
Of the 22 exercises in Learning Perl chapters 1 - 6, only 1 single exercise was omitted because it is not currently supported by RPerl, specifically Chapter 1 Exercise 3, which deals with a purposefully "strange" and "scary" regular expression:
s/\w]+)>/\U$1/g;
The regular expression engine is it's own sub-language and does not require any work from RPerl to be optimized for speed. Initial support for regular expressions will be included in the upcoming RPerl v2.x release series.
...
I hope that helps answer your concerns about a perceived lack of RPerl users. Obviously there are RPerl users, there are RPerl enthusiasts, there are many people who want to see RPerl funded by TPF grants. We're only asking for $1,600 while others routinely request up to $20K at a time. Please support RPerl and vote yes on this and future TPF grant proposals!
Perling,
~ Will the Chill
One more quick note:
RPerl is compatible w/ Perl v5.10 and newer, as specified by the MIN_PERL_VERSION configuration option in Makefile.PL:
https://github.com/wbraswell/rperl/blob/master/Makefile.PL#L20
You can see the RPerl test suite passing on the Travis-CI testing platform:
https://travis-ci.org/wbraswell/rperl
RPerl has been primarily developed on Linux as seen on Travis-CI, and also RPerl has been run on both Windows and Macintosh. We are currently in the process of configuring automated tests for these additional platforms.
Hi,
I'm not anti-RPerl. I'm asking neutral questions. Any project by definition starts with zero users other than its founder - and to get to a large number users it has to pass through a small number of users. What I'm trying to establish is how to measure RPerl's userbase, to see how it changes, not to pass judgement on its size.
At the time of writing, there were no published comments (visible to me or anyone outside of the blog's owners) that stated that they used RPerl. I'm really impressed by the number of people commenting in favour of RPerl. A lot of people have expressed interest. But I stand by my careful distinction between people using it, and people interested in it. There seem to be a large number of potential users here, but not many current users.
Hence seems like a valid reason to improve the documentation, in the hope of getting many of them to try it out (or retry it). But in turn, that's why I think it's very important to be able to estimate the effect on the size and activity of the userbase, to measure the success of such grant-funded work.
I asked about blogs. You cited 5 links to blogs.perl.org. All 5 merely mention RPerl. None state that the blogger has used RPerl
On the subject of multiple Perls, I found it easy to explain why there is a Perl5, a Perl6 and even a RPerl in existence
Will Braswell spoke next about his "Rperl" compiler
which translated his own quite restricted subset (no regexps yet and
no $_) of Perl 5 line by line into C++ in order to run some of the
language shootups benchmarks (a graphical animation of planetary
motion) at increased speed.
This can split various optimizations per module/.pm file, so we can use B::CC compiled modules or even rperl compiled modules, compile-times should go down from 20min to ~5min, with much faster smoker feedbacks, and pushing updates live is much faster, because they will be much smaller.
But let's think, because perl5 has not a offical spec, what if there are many incompatible perl in the world like cperl stableperl and Rperl? Do you really want to stick with perl?
Look at the B::CC and rperl benchmarks.
Nice one +1 from me....
I vote yes. RPerl may make Perl more attractive in general again, and in particular in HPC. Speed is always a good showcase. Also, documenting a proyect forces you to think about, review it, again.
Please yes!
Hi Mr. Clark,
Once the RPerl user documentation is complete, more people will write blog posts about RPerl, and the popularity of Perl itself will increase.
Thanks!
~ Will
Once the RPerl user documentation is complete, more people will write blog posts about RPerl