Leo adoption and the Leo user community

[duf...@gmail.com]

Hi

I am a new user. (actually, I am still testing Leo), and i could not help noticing that the use of Leo does not seem to be very widespread. Very few people talk about it and the user community seems to be quite small.
In terms of user adoption, it seems to me that Leo is nowhere near that of other main text editors. Sure, Leo is in a category of its own, being an outlining editor, as opposed to a simple editor, but should this not make it more popular, instead of more obscure?
I really don't get it.
From what I have seen, Leo has excellent capabilities. And yet it is still a semi-obscure application.
What are the reasons for this? Is it justa a matter of not being publicized enough, or is there something else? Or, let's say, what are the complaints that people have against Leo?

Please, help me to understand this puzzle.

Cheers

[Edward K. Ream]

On Thu, Aug 8, 2013 at 3:25 PM, <duf...@gmail.com> wrote:

Hi

I am a new user. (actually, I am still testing Leo),

​Welcome aboard. Newbies often have the clearest picture of Leo's failings.  All of Leo's core developers take their opinions seriously.
​

 

and i could not help noticing that the use of Leo does not seem to be very widespread.

Very few people talk about it and the user community seems to be quite small.​​​

 

​It's hard to say.  For example, a single download to a professor can, and has, represented​

 

​dozens or hundreds of users.  This reminds me: Leo should encourage users to send "postcards" to us so that we can have a better idea of who is actually using Leo.  Randy Pausch's Alice project (Alice.org) used to do this.  Not sure if it still does.
​

In terms of user adoption, it seems to me that Leo is nowhere near that of other main text editors.

​I think that's right.
​

 

Sure, Leo is in a category of its own, being an outlining editor, as opposed to a simple editor, but should this not make it more popular, instead of more obscure?

​Not necessarily.  People have a large investment in their existing tool chain.  This makes us all reluctant to make big changes.
​

 

I really don't get it.

​ ​

From what I have seen, Leo has excellent capabilities. And yet it is still a semi-obscure application.

 

What are the reasons for this? Is it just a matter of not being publicized enough, or is there something else? Or, let's say, what are the complaints that people have against Leo?

​Leo's core developers have asked these questions often.  It's good to be reminded of them.

It's true: I suck as a publicist ;-)  But there substantive problems as well, especially for teams​:

1. Leo naturally wants to insert sentinels into external files.  This is the only *safe* way to retain outline structure.  Many people, especially those wanting to dis Leo, object to the sentinels.  True, there are alternatives to sentinels (using @auto and @shadow), but they aren't as powerful as @file.

2. Sharing .leo files themselves (rather than sharing the external files) is problematical.  There are workarounds (so-called "ref" files).  See the FAQ entry, How should I use Leo with bzr/git/hg/svn/cvs?, http://leoeditor.com/FAQ.html#id21.

Finally, you should realize that it was never my goal to displace emacs or vim.  My reasons for doing Leo were to create a tool that:

1. allows me to understand complex computer programs.

2. expands the power of scripting.

Those are still my personal focus.  Having said that, it's recently become apparent that documenting what Leo can *already* do has top priority.  Here are the top three items on the list of desired outcomes for Leo:

1. Newbies can install Leo without problems.
2. Newbies can learn Leo's main features easily.
3. *Good* documentation exists for all important plugins and features.

Please, help me to understand this puzzle.

​HTH.  Thanks for your interesting question.

Edward

[dufriz]

On Fri, Aug 9, 2013 at 2:56 PM, Edward K. Ream <edre...@gmail.com> wrote:

 

Very few people talk about it and the user community seems to be quite small.

 

It's hard to say.  For example, a single download to a professor can, and has, represented

 

dozens or hundreds of users.  This reminds me: Leo should encourage users to send "postcards" to us so that we can have a better idea of who is actually using Leo.

Couldn't you include an automatic update feature into the program, which will signal the presence of an instance of Leo being used?

[Jacob Peck]

While possible, it would be akin to spying, and I for one wouldn't want that.  It would need to be opt-in (opt-out isn't good either - lack of consent to the unknowing), and anything like that would be heavily skewed by all the users who don't opt-in.

Additionally, many users run multiple copies of Leo - I have 4 installs myself.

I do think the "postcard" idea that Edward had would be neat though.  Perhaps we could set up a page on the Leo site with all of the postcards/trinkets/etc. recieved?  Could be a neat PR stunt, if nothing else :p

-->Jake

--
You received this message because you are subscribed to the Google Groups "leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+...@googlegroups.com.
To post to this group, send email to leo-e...@googlegroups.com.
Visit this group at http://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/groups/opt_out.

[dufriz]

On Fri, Aug 9, 2013 at 4:18 PM, Jacob Peck <gates...@gmail.com> wrote:

On 8/9/2013 10:13 AM, dufriz wrote:

On Fri, Aug 9, 2013 at 2:56 PM, Edward K. Ream <edre...@gmail.com> wrote:

 

Very few people talk about it and the user community seems to be quite small.

 

It's hard to say.  For example, a single download to a professor can, and has, represented

 

dozens or hundreds of users.  This reminds me: Leo should encourage users to send "postcards" to us so that we can have a better idea of who is actually using Leo.

Couldn't you include an automatic update feature into the program, which will signal the presence of an instance of Leo being used?

While possible, it would be akin to spying, and I for one wouldn't want that.

Spying?!? That would not reveal any private information at all. It would only reveal that one instance of Leo is running.

 

  It would need to be opt-in (opt-out isn't good either - lack of consent to the unknowing), and anything like that would be heavily skewed by all the users who don't opt-in.

Additionally, many users run multiple copies of Leo - I have 4 installs myself.

I do think the "postcard" idea that Edward had would be neat though.  Perhaps we could set up a page on the Leo site with all of the postcards/trinkets/etc. recieved?  Could be a neat PR stunt, if nothing else :p

-->Jake

--
You received this message because you are subscribed to the Google Groups "leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to leo-editor+...@googlegroups.com.
To post to this group, send email to leo-e...@googlegroups.com.
Visit this group at http://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to a topic in the Google Groups "leo-editor" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/leo-editor/svzLY9wqWmQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to leo-editor+...@googlegroups.com.

[Jacob Peck]

On 8/9/2013 10:20 AM, dufriz wrote:

On Fri, Aug 9, 2013 at 4:18 PM, Jacob Peck <gates...@gmail.com> wrote:

On 8/9/2013 10:13 AM, dufriz wrote:

On Fri, Aug 9, 2013 at 2:56 PM, Edward K. Ream <edre...@gmail.com> wrote:

 

Very few people talk about it and the user community seems to be quite small.

 

It's hard to say.  For example, a single download to a professor can, and has, represented

 

dozens or hundreds of users.  This reminds me: Leo should encourage users to send "postcards" to us so that we can have a better idea of who is actually using Leo.

Couldn't you include an automatic update feature into the program, which will signal the presence of an instance of Leo being used?

While possible, it would be akin to spying, and I for one wouldn't want that.

Spying?!? That would not reveal any private information at all. It would only reveal that one instance of Leo is running.

Which is more than some users feel comfortable revealing.  Just saying.

[Terry Brown]

Interesting idea - a good excuse, better than / more likely to be
allowed than a pop-up that just asks for permission to report usage.

Not sure it could be automated - the update installation part - it
would certainly be possible to simply check for newer versions and
report status.

I guess you could check both releases and bzr revision numbers.

Possible output:

Current version: 4.10a / rev 4325
Latest: 4.11 / rev 6321
See http://leoeditor.com/update?ver=410a&rev=4325

or

Current version: 4.11 / rev 6321
Latest: 4.11 / rev 7231
(release up to date, new dev. version avaialble)

http://leoeditor.com/update could be a script which explains the
difference between what the user has (hence the parameters) and what's
out now - although perhaps that's overkill. It could just be a page
which lists current version and explains update options (snapshot,
etc.).

Cheers -Terry

[duf...@gmail.com]

Well, in any case the automatic update feature could be disabled.

I would set it on by default, leaving the user the choice to disable it.

[Terry Brown]

On Fri, 09 Aug 2013 10:18:11 -0400
Jacob Peck <gates...@gmail.com> wrote:

> It would need to be opt-in (opt-out isn't good either - lack of
> consent to the unknowing), and anything like that would be heavily
> skewed by all the users who don't opt-in.

+1 on opt in - could ask the user when ~/.leo is created. In fact...
we could create ~/.leo/myLeoSettings.leo with

README
@settings
@bool check_for_updates = True/False

Where the README node explains how to use ~/.leo/myLeoSettings.leo - we
should be doing that anyway, rather than the current Help ->
myLeoSettings not working unless you create it manually.

Cheers -Terry

[Jacob Peck]

Agreed. I'm okay with the prompt opt-in box. :)

-->Jake

[Terry Brown]

Minor point, if this is done, I guess the implementation will be
something like

rev_info = urllib2.urlopen("http://leoeditor/revinfo.json").read()

Which would need to be in its own thread if we want to avoid an
annoying delay during startup.

Cheers -Terry

[duf...@gmail.com]

What would be a rough estimates of actual Leo users? In the tens, hundreds or thousands?
I know, it involves quite a bit of speculation, but still...

[Dave Loyall]

Subsequent discussion of this matter has focused on finding an acceptable way to receive a ping from new installs.

That's because sending, receiving, and tracking pings are the goals of the new feature.

Little or nothing has been said about how to track the latest version, or how to help the user install the latest version.

That's because those aren't goals of the "automatic update feature".

Therefore, I propose to scrap the "update" part and simply implement a ping feature instead.

Here is a DRAFT:

"""Hello, user. This is a message from the Leo developers.  We have no idea how many people use Leo.  Click here to send us a ping.  The exact content of your transmission will be "HTTP GET /leo/newinstall/linux/python3.14", which as you can see includes information about your OS and python version. and the receiver has been configured to forget your IP address but remember the date and time and the information you sent us.  You can send different information if you like by editing ping.leo, blah blah.  We really appreciate it!"""

This will be easier to implement.  Most of the work will be on the server-side.

It will also be more transparent.  That's a VERY important quality.

Will many folks abstain from participation?  Yes.  That's their choice.  And I think that, unfortunately for us, it TRUMPS our need to know how many people use Leo.

I hope that this helps.

Cheers,

--Dave

[Jacob Peck]

On 8/9/2013 10:54 AM, Dave Loyall wrote:

Subsequent discussion of this matter has focused on finding an acceptable way to receive a ping from new installs.

That's because sending, receiving, and tracking pings are the goals of the new feature.

Little or nothing has been said about how to track the latest version, or how to help the user install the latest version.

That's because those aren't goals of the "automatic update feature".

Therefore, I propose to scrap the "update" part and simply implement a ping feature instead.

Here is a DRAFT:

"""Hello, user. This is a message from the Leo developers.  We have no idea how many people use Leo.  Click here to send us a ping.  The exact content of your transmission will be "HTTP GET /leo/newinstall/linux/python3.14", which as you can see includes information about your OS and python version. and the receiver has been configured to forget your IP address but remember the date and time and the information you sent us.  You can send different information if you like by editing ping.leo, blah blah.  We really appreciate it!"""

This will be easier to implement.  Most of the work will be on the server-side.

It will also be more transparent.  That's a VERY important quality.

Will many folks abstain from participation?  Yes.  That's their choice.  And I think that, unfortunately for us, it TRUMPS our need to know how many people use Leo.

I hope that this helps.

Cheers,

--Dave

I think this is a very good draft.  Makes a lot of sense.  Unfortunately it only captures new installs.  Perhaps it could also check for a key in the database, so it catches existing installs when they update via bzr?  The key would be a simple boolean, whether or not they've seen the dialog.  After that first time, it would never bother them again, regardless of their choice.

-->Jake

On Friday, August 9, 2013 9:13:09 AM UTC-5, dufriz wrote:

Couldn't you include an automatic update feature into the program, which will signal the presence of an instance of Leo being used?

--

[Terry Brown]

My leo email folder has 11,000 messages starting from 2010-08-17.

import mailbox
import email

mb = mailbox.MH("/home/tbrown/Mail/misc/leo")
users = set()
for msg in mb:
users.add(email.utils.parseaddr(msg['From'])[1])
print len(users)

>>> 401

From that we can deduce that the above script generates a set with
401 members :-)

I'm sure user base is always significantly greater than mailing list
participation, but what the ratio would be, I have no idea.

How often is leo-editor-latest.zip downloaded from
http://www.greygreen.org/leo/ ... well, I and only just found this out
now, 220.181.157.*, which is in China, is downloading it constantly, so
it's very hard to tell. Why are the doing that - it makes no sense.
By constantly I mean 67 times in the last 24 hours.

Cheers -Terry

[Jacob Peck]

On 8/9/2013 11:11 AM, Terry Brown wrote:

> How often is leo-editor-latest.zip downloaded from
> http://www.greygreen.org/leo/ ... well, I and only just found this out
> now, 220.181.157.*, which is in China, is downloading it constantly,
> so it's very hard to tell. Why are the doing that - it makes no sense.
> By constantly I mean 67 times in the last 24 hours. Cheers -Terry

... Huh. They're building a Leo supercluster, to script the world!

[Terry Brown]

On Fri, 9 Aug 2013 07:54:49 -0700 (PDT)
Dave Loyall <david....@nebraska.gov> wrote:

> Little or nothing has been said about how to track the latest version, or
> how to help the user install the latest version.
>
> That's because those aren't goals of the "automatic update feature".
>
> Therefore, I propose to scrap the "update" part and simply implement a ping
> feature instead.

I'm sure we'd get much lower opt in with that. You're draft's fine,
but it will get read as

**** ****** ** ***** send information ******** *** ********** ******
** ******** ********** ****** ** *************** ****** ** ********
********** ****** ** ***** ****** ** ******** ********** ******
** *************** ****** ** ******** ********** ******
** *****

and people will click No - that's what I'd do.

If we say "Allow Leo to check for updates?" it's a simpler question,
more people will say yes, and there is the added value of letting
people know about new versions. So, I'd vote for that approach.

Cheers -Terry

[Jacob Peck]

Would we be talking full on release updates, or bzr commits?

I'd like the bzr option myself...

-->Jake

[Terry Brown]

On Fri, 09 Aug 2013 11:33:13 -0400
Jacob Peck <gates...@gmail.com> wrote:

> Would we be talking full on release updates, or bzr commits?
>
> I'd like the bzr option myself...

I think we can do both, although there's a question of how much log
space to take up. The complete information is something like

Your release: 4.10a 2012-05-13
Current release: 4.11 2013-08-22
Your build: 4525 2012-05-13
Current build: 6234 2013-11-18

Just a question of how to format that succinctly.

"1 new release and 535 new development revisions available"... I don't
know, it has to be not confusing for non-programmers.

Maybe

Your build: 4525 (4.10a) 2012-05-13
Latest: 6234 (4.11) 2013-11-18

but when you have the latest release, it would look like:

Your build: 4525 (4.11) 2012-05-13
Latest: 6234 (4.11) 2013-11-18

which will make sense to some and not others. I guess if it's

Your build: 4525 (4.11) 2012-05-13
Latest: 6234 (4.11) 2013-11-18
See: http://leoeditor.com/upgrade

it can be explained in more detail there.

Cheers -Terry

[Matt Wilkie]

Hello, user. This is a message from the Leo developers.  We have no idea how many people use Leo.  Click here to send us a ping.  The exact content of your transmission will be ...

+1 from me, for the wording and general approach.

Not sure about the utility of having this presented on install though. I install programs almost every week just to have a look at what they do, and then remove or never touch them again.

Maybe trigger after x number of starts, or x number of days used.

Another idea: it's become pretty standard for programs to have a "check for new version" feature, usually under Help or About. Just count those.

and, there's the social media stuff: facebook like button, twitter followers, google+, etc.

-matt

[Edward K. Ream]

On Fri, Aug 9, 2013 at 9:13 AM, dufriz <duf...@gmail.com> wrote:

 

Couldn't you include an automatic update feature into the program, which will signal the presence of an instance of Leo being used?

​I don't how that would be done.  I agree with Jacob doing so would be way too intrusive.

Edward

[Edward K. Ream]

On Fri, Aug 9, 2013 at 9:18 AM, Jacob Peck <gates...@gmail.com> wrote:

I do think the "postcard" idea that Edward had would be neat though.  Perhaps we could set up a page on the Leo site with all of the postcards/trinkets/etc.

​received?

​I mentioned Alice.com because at one time Randy Pausch di​d encourage e-postcards.  But now googling that gets to Hallmark ;-)  There may still be a spot on Alice.com for feedback, but I haven't found it yet.

Does anyone know a suitable e-postcard format that we could encourage people to use?  It would be worth some real effort.

Edward

[Edward K. Ream]

On Fri, Aug 9, 2013 at 9:27 AM, Terry Brown <terry_...@yahoo.com> wrote:

 

I guess you could check both releases and bzr revision numbers.

​I would rather not do anything automatic like this.

Edward

[Edward K. Ream]

On Fri, Aug 9, 2013 at 9:34 AM, <duf...@gmail.com> wrote:

What would be a rough estimates of actual Leo users? In the tens, hundreds or thousands?
I know, it involves quite a bit of speculation, but still...

​I would estimate thousands to tens of thousands of users.

There are far fewer members of this group.

Edward

[Matt Wilkie]

In terms of user adoption, it seems to me that Leo is nowhere near that of other main text editors. Sure, Leo is in a category of its own, being an outlining editor, as opposed to a simple editor, but should this not make it more popular, instead of more obscure?

My take on this is, is that Leo is so different from prior experience, is that we just don't "get it" until hanging around and studying for awhile. It looks and behaves like a text editor at first approximation, but as a straight up text editor it's kind of lacklustre. It doesn't have the beauty and smooth user experience of Sublime, or the apparent power and flexibility of Notepad++ or vim/emacs (stress on apparent).

Similarly as an IDE there's not much initial draw, the code hints and tab expansion in something like Eclipse or PyScripter seem far more advanced and extensive. Again, at initial exposure.

My only experience of outlines prior to Leo was from MS Word, so that wasn't a draw. I just didn't, couldn't, appreciate what having that in a text editing environment meant, let alone clones-in-outlines.

I think of this as similar to seeing perspective drawing in art. It's not innate. The mind needs to learn how to interpret the lines and squiggles on flat paper as 3 dimensional objects. however once learned it takes extra effort to see them as mere flatland again. 

Unfortunately it's a lot harder to get the "aha!" of Leo than of interpreting perspective. Or maybe it's just easier to learn it at 10 years old than at 30!

cheers,

-matt

[Ville M. Vainio]

On documenting side: time.to post more on the leo-editor blog :). I think I'll post about full text search next (bigdash stuff)

--