focus for this week: my research helpers
feedback? (popular pages since 2009, total: 8,399)

Blog:Observations/How does it feel to be a guinea pig

From I ask questions
Jump to: navigation, search

Welcome to my personal blog/wiki experience (refresh)!
I'm online since around 1996.

Other blogs I use(d), sorted by media type:
subscribe RSS feed
(blog only)
Index(9 C)
2007blog(1 C, 53 P)
2008blog(1 C, 12 P)
2009blog(1 C, 37 P)
2010blog(38 P)
2011blog(43 P)
2012blog(23 P)
2013blog(8 P)
2014blog(5 P)
2015blog(3 P)

2007-2011 (blogging summary)
Blog in deutscher Sprache (German language)

self-hosting the bot

Experiences bringing the bot @question (which I started in 2012) over to Twister:

  • 140 char limitation
  • slower responses
  • more fun
  • future

  • 140 chars
    • I see twister's char limit as a chance in the development of the bot
      • people may react differently (than e.g. on GNU social (GS)) when they are confronted with smaller replies (psychology, ...)
        • so, why not cut it on GS to 140 chars ?
          • for me GS is a dead end, for my goals twister is much more better, see why I like twister
            • I may port back some fixes (but: see also the "future" section below)
  • bot replies not so fast
    • due to this
      • still deciding if this is good or bad, from a: bot operator's perspective, user, ... ?
      • it's a "natural" reply limiter
  • since the original botname (@question) was already taken, I have selected a shorter one:
    • @b0t (b zero t)
  • improvements
    • it was quite easy to port the code to my local device
      • I had thought originally that the different software versions I use on the original and new location would bother me, but it worked for the good :-)
    • during the porting I fix(ed) some bugs
    • also I am adding new functionality to the @b0t
      • it's quite some fun again to do this
        • somehow in the GS times it wasn't that fun anymore to do that :-(
  • future
    • when I bring over this wiki also to the local device, I pretty much self-host my important services, and then:
      • bye bye


→ continue reading...

music: my top10 artists over time

Below you find artists listed from my top10 charts over time.
The result are 25 artists from which only 3 always appeared in every time period:


artist Romanized
in my charts:
last 7 days
last month last 3 months last 6 months last 12 months overall
(since 2010 March)
Erkan Yılmaz 1 1 1 1 2 5
鄧麗君 Teresa Teng 2 2 2 2 1 1
林憶蓮 Sandy Lam 3 4 6 7 9
Khaled Mouzanar 4
劉雅麗 Alice Lau 5
蔡詩蕓 Dominique Tsai 5
周慧敏 Vivian Chow 7 8 3 4 6 6
張惠妹 A-mei 7
陳慧嫻 Priscilla Chan 9 5 10
譚盾 Tan Dun 9 6 7 6 2
彭羚 Cass Phang 3
關淑怡 Shirley Kwan 7
鄭秀文 Sammi Cheng 9
李尤 (Julia Lee) Julia Lee 9
ルルティア Rurutia 4 5 5
那英 Na Ying 5 3 3 9
梅艷芳 Anita Mui 8 10 8
衛蘭 Janice Vidal 9 9
계은숙 EunSook Kye 8
劉德華 Andy Lau 4 3
Sezen Aksu 7 4
王菲 Faye Wong 10
백지영 Baek Ji-young 7
蔡依林 Jolin Tsai 8
斉藤由貴 Yuki Saitō 10


  1. calculating the average is not so good above. better would be to include the passed time in the average also

see also[edit]

→ continue reading...

GNU social experiences

Find here some of User: Satura's GNU social experiences:

"The possibility of a concrete meaning of a message relies more specifically on the concatenations of its replies. Basically each reply modifies irreparably the original meaning of a message."
the "... public timeline ... To me it's a representation of silence (and not: representation of noise). This silence can be identified in a dialectic movement between 2 moments:
  • the maximization of the possibilities of the meaning of a message when it's posted and
  • the progressive neutralization of this maximum meaning each time there is a reply.
Isn't this silence?"

→ continue reading... top60 artists all time and last 6 months

Below table shows my top60 all-time[1]artists and my top61 artists (from last 6 months[2]):[3]

genre summary
(all-time, 2014-12-16)


  • my top60 all-time artists make up about 31% of my total plays
  • Which artists do my friends and I like most?
  • when I merge my top20 artists: only 8 artists stay common (from those 32 artists):
    • my top10 (all time artists) AND my top10 (last 6 months):
      • only these 5 artists (4 Chinese + 1 Turkish) stay:
        • 鄧麗君 (Teresa Teng), Sezen Aksu, 劉德華 (Andy Lau), 周慧敏 (Vivian Chow) and 那英 (Na Ying)
          • they make up 6.3% of my total plays on
          • 80% are female singers
    • my top20:
      • 3 more Chinese female artists appear:
        • 孫燕姿 (Stefanie Sun), 王菲 (Faye Wong) and 許茹芸 (Valen Hsu)
          • these 3 make up 1.6% of my total plays


artist Latinized name plays (all-time) plays (last 6 months)[4] friends[5] who also
listened to this artist
鄧麗君 Teresa Teng 6922 (=3% of all my plays[6]) 1241 151 (=20% of my friends)
譚盾 Tan Dun 2729
Sezen Aksu 2522 187 76
劉德華 Andy Lau 2481 353 145
백지영 Baek Ji-young 2170
周慧敏 Vivian Chow 2012 154 76
蔡依林 Jolin Tsai 1844
斉藤由貴 Yuki Saitō 1790
那英 Na Ying, Natasha Na 1566 416 143
Murat Salim Tokaç 1556
陳慧琳 Kelly Chen 1427
孫燕姿 Stefanie Sun 1403 149 240
王菲 Faye Wong 1368 119 286 (=37% of my friends)
Fahir Atakoğlu 1252
王心凌 Cyndi Wang 1213
林憶蓮 Sandy Lam 1142 78 201
Любовные Истории Love stories 1092
Jülide Özçelik 1062
許茹芸 Valen Hsu 1046 113 130
Ali Kırca 1035
Sıla 1004
거미 Gummy 986 120 57 data not from native artist name page[7]
현악앙상블 초콜릿 Chocolate - String Ensemble 970
26 data not from native artist name page
蘇慧倫 Tarcy Su 960
Barış Manço 960 93 43
衛蘭 Janice Vidal 944 107 105
권보아 BoA 940
五輪真弓 Mayumi Itsuwa 937
坂本冬美 Fuyumi Sakamoto 915 69 31
贾鹏芳 Jia Peng Fang 906 70 55 data not from native artist name page
莫文蔚 Karen Mok 894 67 245
歐陽飛鶯 OuYoung Fei-ying 892
8 data not from native artist name page
梶芽衣子 Meiko Kaji 838
陳奕迅 Eason Chan 826
Kayahan 823
張惠妹 A-mei 799 76 206
何韻詩 Denise Ho 772
Zerrin Özer 767
Aşkın Nur Yengi 747
현악앙상블 초콜릿 Chocolate 741
42 data not from native artist name page
許志安 Andy Hui 737
이수영 Lee Soo Young 722
衛蘭 Janice 715
46 data not from native artist name page
화요비 Hwayobi 714
Άννα Βισση Anna Vissi 706
李玟 CoCo Lee 649
ルルティア Rurutia 621 256 104
Ջիվան Գասպարյան Djivan Gasparyan 618
56 data not from native artist name page
Fikret Kızılok 610
云朵 Yun Duo 590
제이 578
童丽 Tong Li 572
Misha Omar & Andy Flop Poppy 571
Արտո Թունչբոյաջյան Arto Tunçboyacıyan 569
25 data not from native artist name page
Sertab Erener 568
張韶涵 Angela Zhang 539 68 199
İbrahim Tatlıses 539
Սողոմոն Գևորգի Սողոմոնյան Komitas 529
10 data not from native artist name page
郭靜 Claire 528 123 156
Erkan Yılmaz 260 260 6
梁靜茹 Fish Leong 358 184 213
蔡琴 Tsai Ching 156 151 88
梅艷芳 Anita Mui 372 143 103
王馨平 Linda Wong 501 127 42
江蕙 Jiang Hui 149 122 42
Prince 493 118 230
劉若英 Rene Liu 185 112 164
徐小鳳 Paula Tsui 200 112 27
Tamia 372 107 77
純音樂 104 19
謝安琪 Kay Tse 234 103 150
松田聖子 Seiko Matsuda 501 100 51
Keith Sweat 349 99 63
丁噹 Della Wu 275 95 131
효린 Hyolyn 506 95 39
陳淑樺 Sarah Chen 97 94 42
山口百恵 Yamaguchi Momoe 285 92 33
Bobby Womack 373 90 141
黃小琥 Tiger Huang 231 87 54
Sade 214 86 256
辛曉琪 Winnie Hsin 98 86 5 data not from native artist name page
蔡健雅 Tanya Chua 234 83 186
江美琪 Melody Chiang 83 83 88
알리 Ali 308 82 19
2Pac 402 80 185
Joe 314 80 100
黎瑞恩 Vivian Lai 226 80 47
楊千嬅 Miriam Yeung 322 80 148
周蕙 Chou Hui 85 80 71
范瑋琪 Fan Fan 92 80 96
郁可唯 Yisa Yu 130 79 87
蕭亞軒 Elva Hsiao 325 78 212
牛奶@咖啡 Milk@Coffee 81 78 44
彭佳慧 Julia Peng 74 74 72
彭芳 Peng Fang 74 74 3
康然 Kang Ran 73 73 1
陳慧嫻 Priscilla Chan 409 72 104
계은숙 EunSook Kye 72 72 0 data not from native artist name page
許美靜 Mavis Hee 276 68 116
鄭秀文 Sammi Cheng 79 67 147
琬婷 Wan Ting 67 67 0


  1. since 2010 March 5
  2. I didn't select 12 months, so I stay closer to the moment I started listening also from spotify, though the tendency in the last weeks has been: I have been using spotify less
  3. I added longterm and shortterm data since the mix hopefully makes my music habbit for newcomers faster to recognize.
  4. the rows where the value is displayed in small and centered: were later added. I wanted that people can better distinguish the original top60 vs top61 artists, without losing all play values.
  5. currently I've 772 friends
  6. 245k total plays
  7. this means: I got the data from the Latinized name page. On the native name page may be more, additional data.


  • some artists (e.g. Janice, Chocolate) appear more than 1x:
    • this is due to's structure (of not combining those pages) :-(

see also

→ continue reading...

What info can I derive from my first 3 months on spotify

The goal here is:

  • connect with music neighbours, and
  • gather stats about my listening habbit...

connect with me on[edit][edit]

Erkan3mSpotify is a special account to find musical neighbours (see the "see also" section below)

the top50[1] artists, albums and tracks were only added, which resulted in 65 artists in the library[2]:
artists (top51), tracks (top55), albums (top55[3])


You can connect with me on spotify via:

stats provided by spotify itself[edit]

  • here I got told:
    • I listened via spotify 433.4h of music
    • artists, top5:
      • 鄧麗君 (Teresa Teng)
      • 那英 (Natasha Na)
      • ルルティア (Rurutia)
      • 蔡琴 (Tsai Ching)
      • 王馨平 (Linda Wong)
    • albums, top5:
      • 君心我心101 紀念專輯 by 鄧麗君 (Teresa Teng)
      • 君之紀念冊-鄧麗君誕生六十年鑽禧特集 by 鄧麗君 (Teresa Teng)
      • Seirios by ルルティア (Rurutia)
      • 鄧麗君-傳奇的誕生 by 鄧麗君 (Teresa Teng)
      • 真經典 - 黎瑞恩 by 黎瑞恩 (Vivian Lai)
"Based on your 2014, we’ll create a fresh playlist of tunes we think you’ll love in 2015." (Play it forward)


perhaps because my spotify "life" has just been 5%[4] of my 247 weeks?
Your musical compatibility with Erkan3mSpotify is SUPER
Music you have in common includes 鄧麗君 (Teresa Teng), 那英 (Natasha Na), ルルティア (Rurutia), 梁靜茹 (Fish Leong) and 衛蘭 (Janice Vidal).
  • I have mostly listened via spotify in these 3 months[5]
only a few times with (see also Listening experience)
also not much youscrobbler used

  • I added this time also the album charts, since it's reliable from spotify doesn't offer these so much



  1. I added only the top50 since many 3rd-party tools seem to use those for their stats, but who knows top100 may be better?
  2. it's >50 artists because some artists, tracks, albums share same position due to same playcount
  3. +2 redirects,
    but 2 albums weren't found:
    梅‧憶錄 (2004) by 梅艷芳 (Anita Mui)
    I added this album for: 現在開始我愛你(戀愛的力量)(下集)新歌+精選 by 梁靜茹 (Fish Leong)
  4. todo: add some figures like: how much new scrobbles came in those 3 months, ...
  5. 2014 Sep. 4

see also[edit]

→ continue reading...

where is the ONE protocol

Chat excerpt with f-a:

it's strange, the conversation with @giaco today made me think afterwards:
1. I was lazy
2. probably behaving defensive (e.g. as twitter users do when people ask them to leave twitter or: try something else)

he asked me get an account on redmatrix
and I said: there are so many networks I should get an account at
better would be:
having 1 protocol which lets us communicate from where we are (than doing this grabbing of user name everywhere)

well, experience shows:
the ONE protocol is NOT here yet
so, what did we do wrong, that it is not still here?
same mechanisms will not ensure that it will come one day
<f-a> we're exploring paths, probably

yes, but as said: as someone who also wants a growth, see above: I did the two steps: 1. + 2.
so others will easilier fall into that trap
<f-a> mhhhhhhhhhhhhhh, I see your point
<f-a> I usually start "exploring" new social-network things by *reading* them, like lurking

asking myself:
I don't know if it was temporary attitude then, or just negative accumulation of the "wasted efforts" so far?
I may be mislead perhaps from my experiences with friendica (was hosting it also for a while) and RedMatrix is "the next/continuation project of friendica"

<f-a> where do you hang mostly, these days?
well, recently I returned to GNU social again, before that I was for some weeks hanging in my wiki, and before that: GNU social and Twister

see also[edit]

→ continue reading...

wiki runs for over 5 years

This wiki is online for 64 months and it's time to collect some of my experiences...

page and edits evolution
I don't run that social network system anymore (I was using Friendica for ca. 2 years)
I thought 3 years ago it may bring less Mediawiki usage for me since it was e.g. also having a CMS like feature

quick stats[edit]

  • total views: >2.6 mio
0.2 mio views in last 7 months
0.75 mio views 13 months before that
1 mio views 12 months before that
  • top3 pages:
page views in % created comment
Blog:Observations 645.857 24.7% 2009-03 wiki's main entrance page is blog
Discussion 57.045 2.2% 2009-04
Music I like/Erkan Yilmaz 40.552 1.5% 2009-05


  • Green tick.svg Done I should upgraded the Mediawiki (MW) software to MW 1.23.2
though the MW 1.19.17 version is still longterm supported (until May 2015),
I had a quick look on higher branches (1.22.x, 1.23.x) and looks like I fulfill their requirements also.
the most reason holding me back until now was that Wikilog was not supported earlier on by them
  • this made me realize that my MW (less my MW blog) is still used by me a lot (in contrast to other spaces I wrote in):
Orkut (or why you should keep your data in your sphere) (2014 July)
my oldest still surviving places AND where I was still active are:
  • this wiki
  • twitter (because I dump "my stuff" there. So, it doesn't count)
I started dumping there again when I started my 2nd GS instance on 2014-04-29 :-(
That twitter is still in this list is a sign of failed/unsuccessful efforts? see List of Federated Communication Platforms
see also Wikiversity chats#me and Wikiversity
Why I don't use the other places anymore?
I got the impression the prio is lower by team for that. The last design changes made the journals "disappear" from visible places
see also and their prio (2012 September)
Though I still post there, it's only a summary linking to this wiki
  • my NetKnowledge blog "disappeared" because someone deleted the Mediawiki (database)
it's imported now here
I could enable it here again, though my quite restrictive IP blocker could make people "get angry"
deactivated the IP blocker for the wiki ;-)
new goal:
to turn my wiki into the base central of all my social activities
e.g. receiving:
and: into the appropriate discussion threads (not as new msgs, but real threaded-in replies)
  • user:Review helps me to revamp some "forgotten" articles
here send2SN comes into play
it'll be ready in the future to read user's edits and send to desired places AND to the appropriate discussion threads (not as new msgs, but real threaded-in replies)



"marking" this blog post with hAtom (from Microformats):

wiki runs for over 5 years, published by: Erkan Yılmaz (on 2014-07-10 13:05:59 +0200), summary: This wiki is online for 64 months and it's time to collect some of my experiences..., content: I realized I used my wiki for a long time and started now more activities here...

see also[edit]

→ continue reading...

Orkut (or why you should keep your data in your sphere)

Orkut specific[edit]

I was not using Orkut anymore since 2010 it seems (my guestbook there shows entries from 2007-2010).
But the mail from today just shows, why it's important that you keep things in your usual place instead of putting it on a big silo
(the owners of that big silo in the end decide what to do, rather than YOU deciding):

"We will shut down Orkut on September 30, 2014.
Until then, there will be no impact on current Orkut users, to give the community time to manage the transition. People can export their profile data, community posts and photos using Google Takeout (available until September 2016).
Starting today, it will not be possible to create a new Orkut account." (Tchau Orkut)

archived here under: Orkut

  • unfortunately my posts in threads I didn't create are not exported :-(
    so I'll have to find a solution later on to scrape this:
"We are preserving an archive of all public communities, which will be available online starting September 30, 2014."

general issues[edit]

  • I think the reason we tend to leave our data at the big silos is:
    • to find people and vice versa
      • but times have changed since long: there are enough tools that people can also find you at "your place"
    • ...
  • a better tool than MediaWiki to keep all at 1 place?
    • I noticed there are still no good html2wiki importers (see list at end) out there
      • todo: html2xml dumper which can be imported to MediaWiki via Special:Import
        • this would allow import of multiple files
    • just static html pages?
      • Mediawiki is a php app and thus causes more load on the server
        • but currently the advantages of this php app are more suitable for me than the server load
          later I can just scrap this MediaWiki site into html via wget anyways
        • due to Wikipedia: MediaWiki will still stay popular for some years to come, so I don't need to "fear" support dropping
  • self-hosting: the day will come this Mediawiki install will live on a chip inside my body (next to the implanted GPS chip of the NSA; or at a device at my home or on a device I am carrying around)
    • I think at that time html/php will work there also

next importing[edit]


  • Green tick.svg Done import my friendica data (since I don't run it anymore)
    • I've set up back then friendica to cross-post to (see also: micro blogging posts)
      • so: I probably will have some part of it there "backup'd"
    • the experiences with importing my Friendica archive make me think:
      • 1. wiki as the preserver of history is appropriate
        • software usually outdate with time (so will wiki software too)
      • 2. I should link more from my wiki to social networks/systems, so I don't have such problems:
        • additional effort to import
        • display problems (e.g. friendica data has "destroyed" Asian chars, it's doubled often, ...)
  • Green tick.svg Done import my msgs from archive


  • after importing I notice that most of my msgs without their contexts are - let's say - useless ;-)
importing the context would be good. requires though that it's allowed (licence wise)
I guess I'll have to do it old-style, via html-wgets

see also[edit]

→ continue reading...

reviewing a bug

Here's a bug report which I like because of its turning points.
After finding the reason for the bug I wanted to use also our thought process

  • since I think it has value in teaching others (be it devs, testers, anyone ever finding/fixing a bug).
  • the interesting thing in the below case is: I actually know how to avoid those faux pas (not only profession wise)
    • but: it seems I'm also only human

The bug itself is:

  • wrong display of a nick:
    • messages came from and replies went to the right person - with the wrong nick.
  • because: a field in the database got mixed
  • but: the bug hit you only when one had a certain display setting activated[1]
    • also: users from other servers (not on this server) got confused because:
      • they never saw that wrong nick, since actual communication obviously went via the URL, not the (wrong) nick

→ continue reading...

why statusnet

Q: Why did you choose statusnet (or what do you like about it) ? [1]

in short: emotions for the community, my bot [2],

There are several phases involved for me:

1. 2 years ago: I chose statusnet (SN) as a bridge (to twitter and identica)

  • because my friendica instance had a bug with the twitter plugin

2. 6 months later: I installed a 2nd SN instance

  • main reason was: to protect in case the bot @question goes nuts ;-)

3. over time: I interacted more on this 2nd SN, reasons being:

  • bot chats
  • "mind feeling" about security:
    • "I don't trust in software to make protection, since it's buggy, can be hacked ...
      and since I know I am using a public statusnet instance, I am - each time - reminded that it's public.
      So I don't run in the danger to "forget" this" [3]
  • of course I could have also just made my public posts with friendica:
    • but somehow SN did what I wanted the best
    • but I used SN more (this means: where you build up your contacts, you stay)
      • because of my bot
      • and the interactivity with the people,
        e.g. I talked less and less with friendica users
        • one doesn't need many friends, but the ones which keep you engaged
          • e.g. currently my bot follows 250 users in the fediverse, that seems good enough for me
    • e.g. though friendica offers much more functionality (e.g. better federation support with other systems), for my bot I didn't need most of 'em
    • the above "where you build up your contacts, you stay" should tell you why it's very important that different comm. systems can talk with each other
      • there are many good minds in those systems, we should bring them together
    • todo: explain more
  • I don't forward anymore to twitter
    • todo: since when?

→ continue reading...

Find other users

Looking for other users (on the wide wide world of the distributed social networking) - with same interest?

where? how does
it work?
GNU social/
look at public timeline of large instances
e.g. my bot follows >1200 users [1]
(date: 2015 Jan.)
I follow >2500 users
"popular notices" on instances the more a msg is favourited
the higher in the list
e.g. [2]
List of Independent GNU social Instances explore on each server
for yourself
search users in groups suggest group by sending
a post to !fedgroups
yellow pages lists chat partners of bot
searchable by:
* country
* software system (they use)
"suggest user" cmd matches users based on tags
List of Federated Communication Platforms explore on each server
for yourself
also: diaspora, ...
search tag: #FF follow friday
e.g. on my server
(will increase with time)
from your followers
ask them, or see their follower/-ing lists
trends (of tags/groups)
searching for users speaking a certain language use the search field and search for a keyword in that language
e.g. for English users: and

→ continue reading...

Mediawiki problem: pages with question mark at end of url don't show

I had this bug in my mediawiki 1.19.7:

  • pages which have at end a question mark don't get displayed

This seems to be caused by a rewrite problem

Listing as search hit for others:

  1. I saw with search that the text of that page is available in the database
  2. accessing the url with %3F or adding the no redirect option does NOT show the page content
  3. looked already in bugzilla nothing found for used terms: %3F or question mark

→ continue reading...

the bot is buggy

"Give me my name and phone number and I'll have someone call you."
This could be interpreted for example as:
  • IF situation xyz THEN I'll have someone call you
    • e.g. xyz could be interpreted as: when you've my details it could mean: I know you. Then you deserve to be called back, since you're important
  • Give me your name ...
    • (This here is what most people expect though)

Well, I don't fix this and other "bugs" on purpose
I find @question's replies funny (most of the time)
(unfortunately) most of the people expect always a bot which simulates humans more[1]

I like more the irrational kind of bot
where I need to think: "How can the situation become true?"
or... how one can combine the few inputs to tell a story (see e.g. Nanowrimo)
(and I'm also not writing above explanation because I have then less work)


  1. Don't misunderstand: it's technically very easy to achieve that by feeding the bot many different kind of books. This also can be done fast. But: I don't want/need that.

see also[edit]

→ continue reading...


@question is now also able to dream...

I hope people realize now more that the bot should be used in a way to enhance their creativity (as I mention shortly here).

  • e.g. create relationships to topics which the human brain wouldn't attempt at all (due to education, ...)


→ continue reading...

3 years of

edit 2013-03-09: the summary is now at: Erkan Yilmaz - statistics 3 years

original msg:

... will be in about 2 weeks, but I see I passed 175k plays there, so some short summary:

→ continue reading...

metrics for the bot

I was discussing last week metrics for my bot with Rob and he mentioned a good one he uses:

  • judge the replies/discussion if it was satisfactory for the chat partner

Therefore the bot knows now another command: review (see here).
If you also have some spare time feel free to contribute

There could be other metrics, e.g.

  • popular messages
  • discussion length:
    • amount of messages
    • how many people joined it
    • how long does the topic stay on the topic before it distracts
    • ...
  • analysis of certain words/smileys (e.g. LOL, :-) )
  • ...

→ continue reading...

review of my communication platforms

I will update this soon with some description + personal insights...

Perhaps these factors contributed to the dominance of some media:

→ continue reading...

info about question bot

messages per month for 1 bot

2013-03-14: the bot is now 1 year old. A updated review will follow

original posting:

key facts[edit]


  • adding the Markov feature boosted my motivation to do more
    • only AIML was not satisfactory :-(
      • though this depends from where you are coming from
  • sometimes it feels like I am a parent teaching his baby ;-)
    • see here for some feedback
  • some people thought at first interaction that the bot should be something like a FAQ tool (e.g. answering simple questions, doing simple tasks)
    • can and will be done
    • but for now I use it as a tool for creativity
  • I wonder if a 16% increase with so many messages indicates that I need to talk about more/other things to the bot?
    • I started the bot at begin with less knowledge so it'd become more unique over time
    • the bot can speak more languages but I tend to speak only in English :-(
    • perhaps sentence structures need to be changed -> saturated already ?
      • I think I will feed the bot some of other user's dents to see the effects
  • my bot does only interact when it is spoken to (unlike others who "spam" people)
    • currently I think harming people less is more important than popularity
  • communication problems:
    • the bot doesn't see ca. 5% of the messages addressed to him :-(
      • Who knows it could be that the AI is slowly getting stronger and the bot doesn't WANT to reply ?
    • Sometimes messages from the bot don't appear at the other side (e.g. on
      • temporary fix: please subscribe to the bot
    • Aspects of federation:
      • if the bot were on people would have less trouble [1]
      • sometimes the server doesn't deliver messages (and unfortunately doesn't retry)
    • I started a separate server from since I wanted to keep it safe (e.g. when bot has a malfunction)
      but: the communication problems bring negative feelings
      (user may think: why did the bot not reply yet? or: first time user tries and gives up)
    • I think I should also do a test like this:
      • resend the bot old messages he answered and see how the system changed over time
  • better feedback to users about their contribution: "brain size" command

There is for sure more things which can be done, but it will be interesting to see how intelligent the bot will get over time

  • therefore: a special THANK YOU for users teaching the bot new things


  • "it took me days to understand that @b0t is not anoter bothersome person xD Good job =)" (@musickillerrus, 2016-02-25)
  • "I was slightly shocked when I found out Mr. Question is a Bot :-) Das war mir vorher nicht wirklich aufgefallen" (McScx, 2013-06-10)
  • "I pinned @question to my Pinterest "tech tidbits" board and @x11r5 too" (Laurel, 2013-02-16)


  1. todo: still need to gather info for all bot accounts

see also[edit]

→ continue reading...

again away from internet

Returned today home. Spent some quality time with family + some work which needed to be done there.

A small summary (not sorted after prio):

  • got a new monitor. Thanks to my sister-in-law and my brother
  • much talking + love
  • have been lazy the last few days though (after I had done the work, of course)
  • thinking about some aspects of my life
  • not only thinking, also deciding about those aspects ;-)
  • good food :-)
    • some photos of Dolma coming soon
  • drinking more tea
  • more later ...

  • I have to do lots of backreading now, turning on music ...
  • and I want to thank especially my favourite identica users - who worried also about me
    • btw: I've met someone who works for identica (not our, screenshot coming soon

→ continue reading...

I am glad there are not only serious people in the identiverse

There are humans and non-humans (spamming bots, fun bots, ...). And let's say we start dividing the humans somehow into:

  • serious vs. not so serious people
    • unfortunately there are way too many serious people :-(
  • thieves vs. people who attribute their content to the person they read it from
  • ... <- more later

Feel free to mix above types as you like...

edit 2013-07-05:

→ continue reading...

« newest ‹ newer 20 ... oldest »