www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - GSoC 2017 Ideas!

reply Craig Dillabaugh <craig.dillabaugh gmail.com> writes:
So the ideas page is up for the 2017 GSoC.  Its a bit light on 
content.  Please feel free to use this forum thread to discuss 
any ideas you might have for appropriate projects.

https://wiki.dlang.org/GSOC_2017_Ideas

Cheers

Craig
Jan 14 2017
next sibling parent reply rikki cattermole <rikki cattermole.co.nz> writes:
On 15/01/2017 4:19 AM, Craig Dillabaugh wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on content.
 Please feel free to use this forum thread to discuss any ideas you might
 have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig


This year perhaps we should have something for dub.
Jan 14 2017
parent reply Craig Dillabaugh <craig.dillabaugh gmail.com> writes:
On Saturday, 14 January 2017 at 15:23:19 UTC, rikki cattermole 
wrote:

 On 15/01/2017 4:19 AM, Craig Dillabaugh wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on 
 content.
 Please feel free to use this forum thread to discuss any ideas 
 you might
 have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig


 This year perhaps we should have something for dub.


What did you have in mind, new default language for the config 
file perhaps :o)?
Jan 14 2017
next sibling parent reply Adam D. Ruppe <destructionator gmail.com> writes:
On Saturday, 14 January 2017 at 15:54:46 UTC, Craig Dillabaugh 
wrote:

 What did you have in mind, new default language for the config 
 file perhaps :o)?


The website could use some work. Things that would make me hate 
it less (fyi, I don't actually use dub, but I sometimes try to 
keep packages up for other people):

* search including subpackages

* subpackage navigation (currently we can list them as 
dependencies but that's kinda hideous)

* some kind of documentation scanning or linking, including for 
subpackages. (see, I subpackage everything because I mostly work 
with modules, not repos.)


Things I've seen other people ask for:

* download count

* rating system (btw I think stars aren't that good, I'd rank on 
"did this work for you?" multiple choice: "perfect!", "yes", "no 
it sucked", and "it wasn't what I thought it was". or something 
like that. 1 star might mean the code was awful, or it might mean 
the description was awful. Separating "sucked" with "not what i 
expected" would try to handle that.)



On the program rather than the website:

* integration with external stuff better

* way to install and describe things for offline



So there's a decent amount of low hanging fruit here - I haven't 
looked at the code, but I can't imagine any of this is all that 
difficult to do. The hard part might be an innovative design to 
actually get people to where they need to go more than the code; 
like keeping ratings is easy, just CRUD that number, but sorting 
by rating in search so you enter "website scraping" and it gives 
you back my dom.d (or whatever superior competitor is out there) 
as a high confidence, popular solution front and center instead 
of a list of 0 or 100 things that all look the same to the 
newbie, that's tricky.
Jan 14 2017
parent reply Adam D. Ruppe <destructionator gmail.com> writes:
On Saturday, 14 January 2017 at 16:12:43 UTC, Adam D. Ruppe wrote:

 but sorting by rating in search


I'm sorry, that was a run on sentence.

The big picture goal I'd like to see is that the package manager, 
or even a tutorial author for some topic, just take the choice 
away.

Sure, you can ignore its recommendation and look down the list, 
but it would be nice if you didn't have to; if there was one 
solid way to do what they want that is easy to find. End analysis 
paralysis.

Like on Amazon, where there's hundreds of options, but there's 

the list, it is nice to stop evaluation and just hit buy. It 
kinda sucks to be the new competitor when the system is promoting 


You also want to avoid cheating the system and manipulating the 
results, by either established slumlords or new guys wanting a 
leg up.


If the student can solve the design problem, the implementation 
might be easy. idk how Google would feel about trivial code with 
painful design, but that's the way a lot of software work is in 
the real world sooo I feel it is an applicable project.
Jan 14 2017
parent reply Craig Dillabaugh <craig.dillabaugh gmail.com> writes:
On Saturday, 14 January 2017 at 16:20:06 UTC, Adam D. Ruppe wrote:

 On Saturday, 14 January 2017 at 16:12:43 UTC, Adam D. Ruppe 
 wrote:

 but sorting by rating in search


 I'm sorry, that was a run on sentence.

 The big picture goal I'd like to see is that the package 
 manager, or even a tutorial author for some topic, just take 
 the choice away.

 Sure, you can ignore its recommendation and look down the list, 
 but it would be nice if you didn't have to; if there was one 
 solid way to do what they want that is easy to find. End 
 analysis paralysis.

 Like on Amazon, where there's hundreds of options, but there's 

 the list, it is nice to stop evaluation and just hit buy. It 
 kinda sucks to be the new competitor when the system is 


 You also want to avoid cheating the system and manipulating the 
 results, by either established slumlords or new guys wanting a 
 leg up.


 If the student can solve the design problem, the implementation 
 might be easy. idk how Google would feel about trivial code 
 with painful design, but that's the way a lot of software work 
 is in the real world sooo I feel it is an applicable project.


I've been trying to find something on this, but haven't yet, but 
I am not sure if website work would be considered appropriate.  I 
know pure documentation is not acceptable, and seem to think the 
websites might fall in the same category - but I am not sure. I 
will keep looking.
Jan 14 2017
parent reply Adam D. Ruppe <destructionator gmail.com> writes:
On Sunday, 15 January 2017 at 04:11:06 UTC, Craig Dillabaugh 
wrote:

 I've been trying to find something on this, but haven't yet, 
 but I am not sure if website work would be considered 
 appropriate.


The website is still a program, and the ranking algorithm is, 
well, an algorithm, so it is code; the student would likely be 
researching, experimenting, perhaps gathering data, and 
ultimately, submitting D code with the result.

Though, it wouldn't be a library like most the other accepted 
projects.
Jan 14 2017
parent reply Craig Dillabaugh <craig.dillabaugh gmail.com> writes:
On Sunday, 15 January 2017 at 05:17:10 UTC, Adam D. Ruppe wrote:

 On Sunday, 15 January 2017 at 04:11:06 UTC, Craig Dillabaugh 
 wrote:

 I've been trying to find something on this, but haven't yet, 
 but I am not sure if website work would be considered 
 appropriate.


 The website is still a program, and the ranking algorithm is, 
 well, an algorithm, so it is code; the student would likely be 
 researching, experimenting, perhaps gathering data, and 
 ultimately, submitting D code with the result.

 Though, it wouldn't be a library like most the other accepted 
 projects.


So there does seem to be significant support for something along 
these lines, anyone interested in mentoring this.
Jan 21 2017
parent reply CRAIG DILLABAUGH <craig.dillabaugh gmail.com> writes:
I wanted to ask if anyone knows if there is a way to add links to 
PDF documents to the Wiki.  I want to post some successful past 
proposals, but the Wiki only seems to want to let me upload a 
small number of formats, with PDF not being one of them.
Jan 24 2017
parent rikki cattermole <rikki cattermole.co.nz> writes:
On 25/01/2017 12:17 PM, CRAIG DILLABAUGH wrote:

 I wanted to ask if anyone knows if there is a way to add links to PDF
 documents to the Wiki.  I want to post some successful past proposals,
 but the Wiki only seems to want to let me upload a small number of
 formats, with PDF not being one of them.


Create a new post requesting PDF enabled for upload on the wiki. 
Otherwise you won't get the appropriate responses.
Jan 24 2017
prev sibling next sibling parent Jack Stouffer <jack jackstouffer.com> writes:
On Saturday, 14 January 2017 at 15:54:46 UTC, Craig Dillabaugh 
wrote:

 What did you have in mind, new default language for the config 
 file perhaps :o)?


Here's a list of issues that are considered blocking by the guy 
who's in charge of getting D into Debian's package manager:

https://gist.github.com/ximion/77dda83a9926f892c9a4fa0074d6bf2b#dub
Jan 14 2017
prev sibling parent rikki cattermole <rikki cattermole.co.nz> writes:
On 15/01/2017 4:54 AM, Craig Dillabaugh wrote:

 On Saturday, 14 January 2017 at 15:23:19 UTC, rikki cattermole wrote:

 On 15/01/2017 4:19 AM, Craig Dillabaugh wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on content.
 Please feel free to use this forum thread to discuss any ideas you might
 have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig


 This year perhaps we should have something for dub.


 What did you have in mind, new default language for the config file
 perhaps :o)?


One thing I'd like code.dlang.org to support is other services. E.g. 
arbitrary git and mercurial servers. Sub directory dub files ext.
Jan 14 2017
prev sibling next sibling parent reply Jack Stouffer <jack jackstouffer.com> writes:
On Saturday, 14 January 2017 at 15:19:23 UTC, Craig Dillabaugh 
wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on 
 content.  Please feel free to use this forum thread to discuss 
 any ideas you might have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig


This isn't enough for GSoC, but maybe it can be rolled into other 
website ideas:

It's a pain in the ass to use GDC or LDC when the website only 
shows the current Phobos documentation. It gives no way to go 
back a version, or three.

Currently I have to use archive.org.
Jan 14 2017
parent Iain Buclaw via Digitalmars-d <digitalmars-d puremagic.com> writes:
On 14 January 2017 at 19:44, Jack Stouffer via Digitalmars-d
<digitalmars-d puremagic.com> wrote:

 On Saturday, 14 January 2017 at 15:19:23 UTC, Craig Dillabaugh wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on content.
 Please feel free to use this forum thread to discuss any ideas you might
 have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig



 This isn't enough for GSoC, but maybe it can be rolled into other website
 ideas:

 It's a pain in the ass to use GDC or LDC when the website only shows the
 current Phobos documentation. It gives no way to go back a version, or
 three.

 Currently I have to use archive.org.


Yeah, this is something that readthedocs.org does a lot better at here.
Jan 14 2017
prev sibling parent reply Jack Stouffer <jack jackstouffer.com> writes:
On Saturday, 14 January 2017 at 15:19:23 UTC, Craig Dillabaugh 
wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on 
 content.  Please feel free to use this forum thread to discuss 
 any ideas you might have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig


Fully implementing the C++ STL integration.

According to Walter, everything that's needed in DMD is there. It 
just requires someone who knows the in's and out's of the STL to 
sit down and create the equivalent structs to get the right 
mangling.

I think this would become a killer selling point of D if this 
were to happen.
Jan 14 2017
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 1/15/17 7:43 AM, Jack Stouffer wrote:

 On Saturday, 14 January 2017 at 15:19:23 UTC, Craig Dillabaugh wrote:

 So the ideas page is up for the 2017 GSoC.  Its a bit light on
 content.  Please feel free to use this forum thread to discuss any
 ideas you might have for appropriate projects.

 https://wiki.dlang.org/GSOC_2017_Ideas

 Cheers

 Craig


 Fully implementing the C++ STL integration.

 According to Walter, everything that's needed in DMD is there. It just
 requires someone who knows the in's and out's of the STL to sit down and
 create the equivalent structs to get the right mangling.

 I think this would become a killer selling point of D if this were to
 happen.


We have Alexandru working on that.

BTW stuff from https://wiki.dlang.org/Project_Ideas should be merged 
into GSoC, too.


Andrei
Jan 15 2017