Thread overview
Popularity of Phobos modules/packages
Apr 27, 2021
Berni44
Apr 28, 2021
user1234
Apr 28, 2021
H. S. Teoh
Apr 28, 2021
jmh530
Apr 28, 2021
user1234
Apr 28, 2021
Berni44
April 27, 2021

I wondered about the popularity of Phobos' modules and used a search on GitHub to get a first idea:

I used searches like this: https://github.com/search?l=D&q=std.magic&type=Code

The result looks like this:

module count note
stdio 63,180
conv 47,297
algorithm 43,560
string 34,394
range 28,258
exception 27,217
traits 26,798
array 24,558
typecons 21,696
math 21,597
format 12,742
meta 9,644
datetime 9,101
file 8,272
container 7,723
utf 7,393
functional 6,694
ascii 6,597
random 6,514
numeric 6,255
uni 5,860
experimental 5,386
internal 5,259
path 5,188
socket 4,206
typetuple 4,160 should not be used anymore
regex 3,958
variant 3,837
concurrency 3,785
bigint 3,692
bitmanip 3,660
process 2,908
parallelism 2,748
getopt 2,738
digest 2,139
json 1,909
windows 1,798
base64 856
encoding 828
system 805
net 764
uuid 753
complex 697
zlib 651
stdint 620
uri 570
compiler 443
xml 410 deprecated
mmfile 410
signals 380
mathspecial 363
demangle 358
outbuffer 315
zip 278
csv 236
sumtype 4 new, not yet in stable
April 28, 2021

On Tuesday, 27 April 2021 at 19:54:08 UTC, Berni44 wrote:

>

I wondered about the popularity of Phobos' modules and used a search on GitHub to get a first idea:

I used searches like this: https://github.com/search?l=D&q=std.magic&type=Code

The result looks like this:

module count note
stdio 63,180
... ...
process 2,908
... ...

std.process low use is a surprise to me. It's possible that the most recent uses are hidden by import std;

Also, I suppose that the results dont filter all the phobos forks out ?

April 28, 2021

On Tuesday, 27 April 2021 at 19:54:08 UTC, Berni44 wrote:

>

I wondered about the popularity of Phobos' modules and used a search on GitHub to get a first idea:

I used searches like this: https://github.com/search?l=D&q=std.magic&type=Code

...

I'd like to mention that, as most must have noticed I think, code search on GH includes forks. All phobos forks should be excluded from your results.

April 28, 2021

On Wednesday, 28 April 2021 at 09:24:47 UTC, user1234 wrote:

>

I'd like to mention that, as most must have noticed I think, code search on GH includes forks. All phobos forks should be excluded from your results.

I'm aware of that. I came upon this, when I tried to estimate how many people use std.signals. I went through all 380 hits and came up with 35 "real" hits. Several of them with their last update years ago.

Anyway, as a first idea, which modules are popular and which not, this should be sufficient.

April 28, 2021
On Wed, Apr 28, 2021 at 07:52:21AM +0000, user1234 via Digitalmars-d wrote: [...]
> std.process low use is a surprise to me.

Yeah, I use std.process all the time; its API is newer (a few years ago it was redone -- in an excellent way IMO) and very convenient. I'd have expected that it would be pretty popular among D users.


T

-- 
Don't drink and derive. Alcohol and algebra don't mix.
April 28, 2021
On Wednesday, 28 April 2021 at 12:58:54 UTC, H. S. Teoh wrote:
> On Wed, Apr 28, 2021 at 07:52:21AM +0000, user1234 via Digitalmars-d wrote: [...]
>> std.process low use is a surprise to me.
>
> Yeah, I use std.process all the time; its API is newer (a few years ago it was redone -- in an excellent way IMO) and very convenient. I'd have expected that it would be pretty popular among D users.
>
>
> T

Of course, low use (on github) does not mean something is worth killing, by itself. std.process is very useful when you need it.
April 28, 2021

On Tuesday, 27 April 2021 at 19:54:08 UTC, Berni44 wrote:

>

I wondered about the popularity of Phobos' modules and used a search on GitHub to get a first idea:

I used searches like this: https://github.com/search?l=D&q=std.magic&type=Code

Did you constrain it by date?

https://github.com/search?q=language%3AD&type=Repositories
https://github.com/search?q=created%3A%3C2019+language%3AD&type=Repositories

Gives 7554 repos of which 6657 repos were created before 2019.

So basically 88% of all D-repos are old.

April 28, 2021

On Wednesday, 28 April 2021 at 14:00:13 UTC, Ola Fosheim Grøstad wrote:

>

On Tuesday, 27 April 2021 at 19:54:08 UTC, Berni44 wrote:

>

I wondered about the popularity of Phobos' modules and used a search on GitHub to get a first idea:

I used searches like this: https://github.com/search?l=D&q=std.magic&type=Code

Did you constrain it by date?

https://github.com/search?q=language%3AD&type=Repositories
https://github.com/search?q=created%3A%3C2019+language%3AD&type=Repositories

Gives 7554 repos of which 6657 repos were created before 2019.

So basically 88% of all D-repos are old.

Maybe created is not so good, perhaps better to use pushed:

https://github.com/search?q=pushed%3A%3E2019-04-28+language%3AD&type=Repositories&ref=advsearch&l=D&l=

Which gives 1131 repos, so 15% of the github repos were pushed to in the past year, so 85% are more or less dead?

April 28, 2021
On Wednesday, 28 April 2021 at 14:05:33 UTC, Ola Fosheim Grøstad wrote:
> Which gives 1131 repos, so 15% of the github repos were pushed to in the past year, so 85% are more or less dead?

Typo, I meant in the past 2 years. One year would be a bit short to estimate whether something has been abandoned or not.

Just for fun, number of repos created in the past year for comparable languages, it gives different results for each search so I ran it 3 times for each:

Zig:       400-   700
Crystal:   700-  1000
D:         800-  1000
Nim:      1200-  1300
Rust:    40000- 60000
Go:     140000-160000
C++:    250000-290000


Using:

https://github.com/search?q=created%3A%3E2020-04-28+language%3AD&type=Repositories&ref=advsearch