How Developers Stay Current Using Twitter

Software Engineering at the Speed of Light:
How Developers Stay Current Using Twitter
Leif Singer
Fernando Figueira Filho
Margaret-Anne Storey
@LSinger @FFilho_ @MargaretStorey

3
#icse14 #twitter4se @lsinger

Why & how do developers use Twitter?
What value does it provide?
What challenges does it pose?
How do developers cope with these challenges?
Why do other developers not use Twitter for work?
4

Related Work
In Software Engineering:
Analysis of sampled tweet contents [Bougie et al. 2011]
Classiﬁcation of tweets [Tian et al. 2012]
Qualitative investigation of Drupal’s Twitter use [Wang et al. 2013]
6
But how and why do developers tweet?

Study Design
Exploratory Survey: get a feel for attitudes
• emailed 1,160 active GitHub users – 271 responded
Interviews: deeper insights about value, challenges
• 27 – from exploratory survey
Analysis: Grounded Theory
• codes, categories, themes, memos, …
Validation Survey: do developers agree?
• emailed 10,000 active GitHub users – 1,413 responded
8

1,413 Validation Survey Participants
10
ave enough time to
—even though they
nnect with Western
m our exploratory
oor match for their
their friends or col-
context, with a low
es, preferably with
e multiple clicks to
iated the brief and
ur exploratory sur-
his limitation. The
that many posters
cifically mentioned
o contain anything
e link-shorteners in
ctice.” [E29]
ons.
nversation. I don’t
two or three people
... it’s nice. And I
very much”. [P4]
survey and interviews.
1,413 GitHub users answered our validation survey. Fig. 1
provides some details on the demographics of these respon-
dents. All except one respondent said they develop software
in some capacity, with the majority (81%) being professional
software developers. Roughly two-thirds of our respondents
said they use Twitter at least once a week (adopters). 70%
wanted to be notified of the results of our study.
81%
19%
67%
33%
70%
30%
(a) (b) (c)
Figure 1: Validation survey respondents.
(a) 81% (1,145) professional, 19% (267) non-
professional developers, 1 non-developer;
(b) 67% (940) use Twitter at least once a week,
33% (473) do not; (c) 70% (993) were interested in
study results.
A large part of the survey consisted of statements, each
related to a theme from our findings. For these statements,
we measured agreement using a five-point Likert-type scale.
6
https:/ / docs. google. com/ forms/ d/
1 FxuT3 ORoSjMqQks7hF2 gRyncjfJ8D0 QBXJaqfmy97pU/
(a) 81% / 1,145 professional developers
(b) 67% / 940 use Twitter at least once a week
(c) 70% / 993 interested in results

Results
Value
• Awareness
• Learning
• Relationships
Challenges & Coping Strategies
Non-adoption
11

Value: Awareness
13
‣ Following developers and projects
‣ Following thought leaders
‣ Promoting project activities
‣ Just-in-time awareness
“I think the main advantage is to be in contact with
people who are developing things that I use. There are
people developing libraries and I can communicate
with them through Twitter.” [P11]

14
Value: Awareness
opers follow other developers, projects, news curators, and
thought leaders. This allows them to stay aware of new
practices and resources in a timely manner and provides
them with access to diverse opinions. Developers also pro-
mote their projects and activities, which may in turn help
the dissemination of knowledge and increase the adoption of
practices and tools.
(1) On Twitter, I follow leaders in my technological niche, which
helps me stay current about the latest technologies and practices.
13% 71%
130 developers 673 developers
(2) Twitter helps me promote projects and technologies I work with.
20% 58%
(3) Twitter helps me keep up to date about technologies and tools I
use for software development.
14% 69%
(4) Twitter helps me stay aware of new trends and practices in soft-
ware development.
(1) On Twitter, I follow leaders in my technological niche, which
helps me stay current about the latest technologies and practices.
13% 71%
(2) Twitter helps me promote projects and technologies I work with.
20% 58%
(3) Twitter helps me keep up to date about technologies and tools I
use for software development.
14% 69%
(4) Twitter helps me stay aware of new trends and practices in soft-
ware development.
17% 65%
strongly disagree disagree
strongly agree agree
Figure 1: Results from our validation survey.

Value: Learning
16
‣ Extending knowledge
‣ Learning serendipitously

17
Kelly Sommers
@kellabyte
15
RETWEETS
29
FAVORITES
Twitter is like the best study group I've ever
had.
Reply Retweet Favorited
1:21 AM - 28 Dec 13
Following
More
Reply to @kellabyte
mrb @mrb_bk
@kellabyte same!
28 Dec
Details Reply Retweet Favorite More
Scott Saad @saadware
@kellabyte Curious how you use twitter as study group. You
prefer it over community sites like Stackoverflow? Or as a
supplement?
28 Dec
Kelly Sommers @kellabyte
@saadware I think they are for different purposes. Twitter
seems very social and organic in how conversations flow
and ppl just jump in
28 Dec
Kelly Sommers @kellabyte
@saadware I’ve had the privilege to see such amazing
people jump into a random technical debate and bring so
much value.
28 Dec
Corey Donohoe @atmos
@mrb_bk you guys are doing it right. !
28 Dec
Home Connect Discover Me Search
78,60978,609
TWEETSTWEETS
286286
FOLLOWINGFOLLOWING
16,38016,380
FOLLOWERSFOLLOWERS
Kelly Sommers
@kellabyte
DataStax MVP for Apache Cassandra and Windows Azure MVP,
Mobile dev maven, backend brat, big data and distributed diva,
Relentless learner. I void warranties.
Canada · kellabyte.com
Follow

Value: Learning
18
ing in conversations helps them learn. Developers said that
the qualities and constraints of Twitter enabled serendipi-
tous, undirected learning, sometimes giving them access to
resources they wouldn’t have been able to ﬁnd themselves.
While participants viewed learning as an investment, they
also thought it was fun and rewarding.
(5) Twitter helps me extend my knowledge by exposing me to tech-
nologies and practices I should learn in software development.
18% 60%
(6) Twitter has helped me become a better programmer.
39% 31%
(7) Twitter helps me learn about things I wasn’t actively looking for.
11% 72%
resources they wouldn’t have been able to ﬁnd themselves.
While participants viewed learning as an investment, they
also thought it was fun and rewarding.
(5) Twitter helps me extend my knowledge by exposing me to tech-
nologies and practices I should learn in software development.
18% 60%
(6) Twitter has helped me become a better programmer.
39% 31%
(7) Twitter helps me learn about things I wasn’t actively looking for.
11% 72%
Figure 1: Results from our validation survey.Figure 3: Results from our validation survey (RQ2)
for developers who use Twitter at least once a week.

Value: Relationships
20
‣ Managing one’s image
‣ Building community
‣ Discovery of interesting developers
‣ Achieving trust and rapport
‣ Work validation and feedback on projects
‣ Job opportunities
(12) Twitter helps me build trust or rapport with other developers.
23% 49%
(13) Twitter helps me receive validation from others for my work.
35% 34%
(14) Twitter helps me give and get feedback about projects I work
with.
30% 44%
(15) Twitter helps me access job opportunities.
44% 28%

Challenge:
Maintaining a Relevant Network
22
‣ value of Twitter is in the network
‣ can become obsolete – move, new job,
different technologies, …
‣ can become too large – overwhelming
“When you follow 1,000 accounts, many things you see
are just not for you.” [P16]

Challenge:
Consuming Content
23
‣ lots of content to consume
‣ Twitter can impact productivity by leading to distractions
(16) I carefully decide whom I follow in order to avoid information
overload on my Twitter feed.
12% 72%
(17) On Twitter, I usually follow people on a trial basis and unfollow
them if they post irrelevant or too much content.
17% 67%
(18) I ﬁnd it hard to cope with the amount of information I receive
on Twitter.
34% 41%
(19) I appreciate the succinctness of 140 characters per post on Twit-
ter.
11% 68%
(20) Twitter is ﬁne for short discussions.

Strategy:
Following Relevant Developers
25
‣ find key developers from one’s niche, follow them
‣ find similar people mentioned by key developers,
consider following them
‣ look at source code before deciding to follow
‣ real-life recommendations from co-workers
‣ look for geographically close developers
‣ read bio for first impression; #followers, #tweets

Strategy:
Unfollowing Developers on Twitter
26
‣ own and others’ interests change – continuously adapt
network
‣ following on a trial basis: relevant? volume OK?
‣ can always unfollow
‣ routines for purging following lists

27
d
d
d
%
e-
e-
is
ve
ht
we
at
e
n
n-
e-
developers told us that this helps them manage the vol-
ume of posts they have to skim. A few developers said
that on Google+, which has no such limit, skimming was
much harder. Therefore, they were more likely to feel over-
whelmed.
800
400
00!
200!
400!
600!
800!
discuss:'adopters'
Legend
Twitter Blogs In Person
Email Chat Other
Google+ Facebook
Figure 6: The channels Twitter adopters said they
use for longer discussions.
“Oﬄine” Discussions

Barriers
29
‣ Too much noise
‣ Peer adoption
‣ 140 character constraint
‣ Poor support for conversations
‣ Unsure of the beneﬁts
“I don’t understand it and I don’t see any purpose for
it.” [E43]

not generalizable to
• corporate settings
• other cultures
• other microblogging tools
• perhaps not even to other active GitHub users!
but: validation survey was encouraging
31
Limitations

Future Work
influence on practice & tool diffusion
challenges of appropriation
• SE-specific tools? how it fits with other tools?
• different personalities, contexts, needs, …
non-adoption difficult to study
other cultures...
33
Western • Educated • Industrialized • Rich • Democratic

@LSinger @FFilho_ @MargaretStorey
Twitter provides value to professional developers
• awareness, learning, relationships
• appropriated, not made for this – problem?
Found challenges & strategies
• our readers agree: helpful!
Non-adoption interesting, difﬁcult to investigate
Developers are an appreciative audience
Do you use Twitter to support your research?
If yes, tell us how by using #twitter4ser

109 retweets. 42 favorites. 270 tweets from others. 38 new followers.
Leif Singer
@lsinger
109
RETWEETS
42
FAVORITES
How Software Developers Use Twitter: study
results are in! to.leif.me/how-software-d…
(HN: news.ycombinator.com/item?
id=6802578)
Reply Delete Favorite
6:10 PM - 26 Nov 13
More
Reply to @lsinger
Ramon Roche @mrpollo
@lsinger awesome read man, glad I could be part of it,
26 Nov
Home Connect Discover Me Search

Comments on the blog post.
WHAT'S THIS?ALSO ON LEIF SINGER'S BLOG
ICSE 2013: Four Days of San Francisco
2 comments • 17 days ago
Leif Singer — Thanks Olga, and thanks for the
company. :)
On Testing Culture in GitHub Projects
3 comments • 16 days ago
Robert Rouse — Not necessarily. Testing what your
application does when, say, the input is longer than
the field length in the database is valid if the …
Reply
• Edit • Reply •
Leif Singer • a day agoMod LukasEder
That looks great, thanks for sharing! Reminds me of Daniel Jalkut's "Elements of
Twitter Style" -- could be helpful for some as well: http://www.red-
sweater.com/blo...
• Reply •
HackBoy • 3 days ago
Glad to have helped you with this research :D
• Reply •
Luis • 3 days ago
And now you have a new follower on twitter. Great post!
• Reply •
realloc • 3 days ago
Thanks for sharing!
• Reply •
behemoth • 3 days ago
great work,really really interesting
• Reply •
Kenneth Kinyanjui • 3 days ago
I am proud to have been part of the research. This will be surely some important info to us especially
as developers and how we interact and continue with development
• Reply •
Duane Johnson • 3 days ago
Glad to have followed your tweet here! Thanks making all of your research public and available to
developers!
5
Subscribe Add Disqus to your site
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
• Reply •
• Reply •
david karapetyan • 3 days ago
For all the devs that get their tech fix from twitter I suggest looking somewhere else like ruby weekly,
javascript weekly, ..., X weekly and also reddit.com/r/programming, reddit.com/r/ruby, etc.
Leif Singer • 2 days agoMod david karapetyan
In another study we're doing we found especially the weekly newsletters to be an important
source to be right now. Email's having a comeback, who would've thought? I like it.
1
• Reply •
r0ash • 3 days ago
Difficult to being social generally is another reason for some developers, not using Twitter or fb etc.
Leif Singer • 2 days agoMod r0ash
I heard that in another study before, actually. Personally, I think "acting socialable" is something
you can practice and get better at. So why not start on Twitter, Reddit, IRC, etc. where you can
hide behind a disposable identity? You can always start over with a new one.
• Reply •
r0ash • 9 hours agoLeif Singer
Thanks, completely agree, this habit could be acquired, all one have to do is to find what
fascinates him/her and what content/topic to discuss. Just started once again :) but with
real identity.
1
• Reply •
LukasEder • 3 days ago
I participated in that study and I'm glad to see that the general opinion matches my own experience.
Twitter is really both an awesome source of niche information, fun, and also a great channel to promote
my own information off my blog (http://blog.jooq.org). It is a great place to connect with key players in
my industry, which would otherwise be out of reach (I've talked to Eric Meijer, Arun Gupta, and other
luminaries!)
But I'd like to point out that I have friends who are in the sports industry, and they feel the same way
about Twitter. These findings really aren't innate to Twitter users who are developers.
1
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
seem to be likable. People like likable people. That doesn't mean you have to agree with
everything -- just have some basic manners and provide potential value (see above) to others.
- Same goes for your tweets (and possibly blog posts) -- be authentic, have manners, share
what you're learning.
- Finally, you could try posting your questions under hashtags. Some communities are known to
monitor questions about their niche and answer those. IIRC the Drupal paper we cite in our
report said something like that.
Most of the above of course is completely unscientific and just how I experience Twitter. YMMV.
• Reply •
whatever • 2 days agoLeif Singer
I think there's some good advice on here on how to maximize success at the Twitter
Game, especially the use of a profile image and bio, but it's also why I typically use
google and ask questions as stackexchange, expertsexexchange, reddit, irc, and seek
out specialized forums before asking anything on twitter.
(Of course there are downsides to the profile image and bio and that's the obnoxious
behaviors of recruiters and interviewers. It makes it scary to ask a dumb question on the
net, or even a newbie question.)
Leif Singer • 2 days agoMod whatever
Yes, why not use other channels if they're better suited for the question you
have? In our study, Twitter seemed to be much more valuable for serendipitious /
undirected learning than for needing an answer to a specific problem right now.
• Reply •
AsherBond • 2 days ago
I was asked to participate in community research and when the results and findings were shared I also
shared these findings with my followers. A good leader is a good follower and I think thought leaders
are good at following Technology which leads the software industry.
Leif Singer • 2 days agoMod AsherBond
Thanks Asher! I'm blown away by all the tweets, retweets, and favorites this blog post got. Very
thankful.
whatever • 2 days ago
So I have followed 30 dozen thought leaders and 1200 dozen of the people in their networks. But only
16 people follow me, half of those spammers.
So just how do I ask a question? #PissingIntoTheWind
Share ›
Share ›
Share ›
Share ›
Share ›
sports industry
Likewise super interesting. Our study was focused on developers of course, but there are tons
of other publications where people studied Twitter (users) in different contexts. It could be neat
to compare findings across fields, actually.
• Reply •
see more
LukasEder • 2 days agoLeif Singer
> would you care to chime in the comment on asking questions / talking to thought
leaders above?
I'm not sure if I understood this. You want me to link to the discussions that I've had?
Here's the one with Eric Meijer: https://twitter.com/headintheb...
tech.pro/blog/1689/does… Txs ;-) Java 8, Scala, Clojure,
...are why I jumped ship. The JVM is the stable thus
smart choice for the enterprise.
5:41 AM - 2 Nov 2013
Does Java 8 Still Need LINQ? Or is it Better than LINQ?
By jOOQ @JavaOOQ
LINQ was one of the best things that happened to the .NET software engineering
ecosystem in a long time. With its introduction of lambda expressions and...
Erik Meijer
@headinthebox
FollowFollow
Leif Singer • a day agoMod LukasEder
You want me to link to the discussions that I've had?
Well, I basically wanted to know whether you had any hints on how to approach /
get the attention of famous people, because @whatever had brought that up
above. So you wrote a great blog post that caught the attention of Erik -- that's
awesome, and I think is pretty much in line with what I thought might help: put
out great content that others find interesting. If you're a novice, tell others what
you're learning.
• Reply •
LukasEder • a day agoLeif Singer
Ah, well, I'm generally following Alex Turnbull's advice here:
http://groovehq.com/blog/1000-...
Essentially, it is about engaging without spamming
Share ›
Share ›
Share ›
Share ›
• Reply •
• Reply •
david karapetyan • 3 days ago
1
• Reply •
• Reply •
real identity.
1
• Reply •
luminaries!)
1
Leif Singer • 2 days agoMod LukasEder
great place to connect with key players
Interesting -- would you care to chime in the comment on asking questions / talking to thought
leaders above?
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
follow over time.
- Have a profile image and bio that shows you're an actual human being. You don't need to
reveal your identity / gender / etc. but anything beyond the default image helps. It helps if you
• Reply •
• Reply •
thankful.
Share ›
Share ›
Share ›
Share ›
Share ›
• Reply •
1
• Reply •
• Reply •
real identity.
1
• Reply •
luminaries!)
1
Leif Singer • 2 days agoMod LukasEder
great place to connect with key players
Interesting -- would you care to chime in the comment on asking questions / talking to thought
leaders above?
Share ›
Share ›
Share ›
Share ›
Share ›
Share ›
follow over time.
- Have a profile image and bio that shows you're an actual human being. You don't need to
reveal your identity / gender / etc. but anything beyond the default image helps. It helps if you
• Reply •
• Reply •
thankful.
Share ›
Share ›
Share ›
Share ›
Share ›

Hector Zarate (@iOSCowboy) blogs about our blog post.
Tweeting Developers Published 27 Nov 2013
For a long time I saw Twitter as a source of mild entertainment which occasionally brought me
up to update on something relevant.
But that point of view changed recently, when I cherry-picked who I'm following and
discovered the interaction I can have with some great developers.
In no time it became a valuable tool that updates me on what's happening in my tech niche,
tips me about new interesting things to learn and puts me in touch with other cool people
doing cool stuﬀ.
But don't take my word for granted. Leif Singer (@lsinger), a mad scientist from the University
of Victoria in Canada, made a thorough research on How Software Developers Use Twitter.
He wrote a great post about it on his blog that I just can't recommend enough. It includes a
tl;dr and usage strategies for developers using Twitter. Pure gold.
Check it out. You know, for science!

YES, WE DO HAVE BACKUP SLIDES!

41
Staying Aware Spreading Information Managing Relationships
Twitter Adopters
800
400
00!
200!
400!
600!
800!
stay%aware:%adopters%
800
400
00!
200!
400!
600!
800!
spread:(adopters(
800
400
00!
200!
400!
600!
800!
rel:%adopters%
Twitter Non-Adopters
400
200
00!
100!
200!
300!
400!
stay%aware:%non,adopters%
400
200
00!
100!
200!
300!
400!
spread:(non+adopters(
400
200
00!
100!
200!
300!
400!
rel:%non(adopters%
Legend
Twitter Google+ Chat Facebook Other
Email Blogs News Aggregators In Person
igure 7: The channels Twitter adopters and non-adopters use for di↵erent activities

Exploratory Survey Participants: GitHub
42
NA: 100
EU: 69
NO: 49
SA: 13
AU: 3
CA: 5
WA: 7
MA: 5
EA: 3
AF: 0
hirable: 80
not hirable: 171
public_repos followers following followers/following created_at updated_at public_gists contributions
MEDIAN 15 7 5 0.857142857142857 2011-01-27 21:56:32 UTC 2013-09-06 11:08:00 UTC 2 226
MIN 0 0 0 0 2008-01-30 23:19:18 UTC 2013-06-06 11:37:00 UTC 0 0
MAX 301 1321 621 456 2013-05-27 14:43:57 UTC 2013-09-07 07:16:27 UTC 235 57875
AVG 27 40 23 5 2010-12-02 18:11:37 UTC 2013-09-02 11:37:34 UTC 12 814
STDEV: 35 134 57 33 503d 12h 2m 7s 627ms 10d 23h 0m 40s 720ms 26 3785

Exploratory Survey Participants: Twitter
43
Since Last tweeted at Tweets Favorites Listed Following Followers followers/following
MEDIAN 2009-05-04 09:03:37
+0000
2013-09-05 13:10:45
+0000
1124 19 5 169 144 1
MIN 2006-10-26 14:04:33
+0000
2008-01-30 13:57:31
+0000
0 0 0 0 0 0
MAX 2013-05-06 15:40:39
+0000
2013-09-07 05:36:12
+0000
41079 9107 949 2600 11469 125
AVG 2009-07-08 06:44:04
+0000
2013-07-25 03:03:33
+0000
3356 251 32 287 490 3
STDEV: 526d 20h 54m 18s
872ms
185d 23h 19m 18s
564ms
6398 882 103 369 1218 10

Interviewees: GitHub
44
public_repos followers following followers/following created_at updated_at public_gists contributions
MEDIAN 25 13 11 2 2010-08-10 22:10:48 UTC 2013-09-06 17:50:24 UTC 5 385
MIN 2 0 0 0 2008-02-02 23:25:39 UTC 2013-09-03 23:19:51 UTC 0 18
MAX 133 936 162 23 2013-04-24 21:22:13 UTC 2013-09-07 07:02:06 UTC 88 3068
AVG 38 72 21 4 2010-06-15 11:05:34 UTC 2013-09-06 08:44:21 UTC 13 728
STDEV: 37 189 32 6 488d 10h 5m 44s 457ms 0d 22h 6m 31s 167ms 19 878
NA: 12
EU: 8
NO: 1
SA: 3
AU: 1
CA: 0
WA: 0
MA: 1
EA: 1
hirable: 9
not hirable: 18

Interviewees: Twitter
45
Since Last tweeted at Tweets Favorites Listed Following Followers followers/following
MEDIAN 2009-04-02 15:13:28
+0000
2013-09-06 15:28:18 +0000 2210.5 18.5 7.5 165.0 195.5 0.969495091164095
MIN 2007-07-11 00:01:42
+0000
2013-07-24 13:07:57 +0000 4.0 0.0 0.0 23.0 11.0 0.162878787878788
MAX 2012-02-10 20:12:06
+0000
2013-09-07 05:35:37 +0000 29644.0 2101.0 739.0 1999.0 7921.0 19.8
AVG 2009-03-19 22:57:10
+0000
2013-09-02 13:10:19 +0000 4983.1 129.2 55.4 337.0 716.4 2.8
STDEV: 418d 1h 26m 3s 456ms 11d 0h 58m 3s 698ms 7139.5 410.2 153.1 461.5 1635.4 4.4

Validation Survey Participants: GitHub
46
hirable: 482
not hirable: 931
bio public_repos followers following created_at updated_at public_gists contributions
median 17 8 7 2011-01-20 10:36:26
UTC
2013-11-09 03:45:03
UTC
2 203
min 0 0 0 2008-01-12 16:46:24
UTC
2013-05-31 19:08:18
UTC
0 0
max 742 1191 1338 2013-11-07 00:07:11
UTC
2013-11-11 04:05:47
UTC
1549 8159
avg 28 28 22 2010-12-18 11:52:17
UTC
2013-11-05 11:13:12
UTC
15 450
stdev 43 71 58 501d 8h 54m 19s
425ms
14d 14h 16m 22s
356ms
57 686

Validation Survey Participants: Twitter
47
Since Last tweeted at Tweets Favorites Listed Following Followers
MEDIAN 1657d 4h 11m 24s 23d 21h 38m 51s 1087 26 185 154
MIN 2658d 23h 10m 7s 2055d 21h 44m 56s 0 0 0 0
MAX 55d 9h 32m 18s 21d 20h 34m 38s 155210 48918 4056 31678
AVG 1551d 15h 54m 19s 972ms 46d 11h 9m 0s 414ms 3982 387 307 519
STDEV: 577d 19h 8m 52s 816ms 123d 5h 34m 50s 731ms 9988 2250 411 1782

How Developers Stay Current Using Twitter

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (20)

Similar to How Developers Stay Current Using Twitter

Similar to How Developers Stay Current Using Twitter (20)

More from Margaret-Anne Storey

More from Margaret-Anne Storey (10)

Recently uploaded

Recently uploaded (20)

How Developers Stay Current Using Twitter