adoption takes a while, when it happens

Discussion of technology related topics including news, events, and announcements.

adoption takes a while, when it happens

Postby fig1 » November 20th, 2015, 7:47 am

this is a reply to edward, author of netman for devuan-- a person i hold in high regard.

he asks,
Is netman being adopted or was it a waste of time and effort?


i may be a netman user at some point, and im sure steve litt or someone will make the following points, but i feel i should say:

its going to take a while. really-- and the reason isnt even bad.

first of all, most people arent even using devuan (im using it to type this, but thats beside the point.) so all the work thats gone into devuan, was it worthwhile? i certainly hope so. i would be operating-systemless without devuan right now-- ive even tried bsd.

i use some equpiment refracta wont run on. i dont really want to go back to non-debian-ish distros, ive tried about 50 of those and every positive step ive taken in 10 years was towards debian and then devuan.

but how is devuan adoption going? well, its slow-going.

but more people adopt things as time goes on. i was just watching a video about how programming languages are chosen yesterday.

you think its about performance and good design, right? well, according to this data: http://i.imgur.com/vAdlhgU.png

* network effects over performance
(adoption begets adoption, and people use what theyre using)
* peformance over safety (we know facebook wasnt written in ada)
* some "particular language feature" over "simplicity"

these are general data, and there are lots of smaller-scale exceptions.

but the top item is that many more people (tend to) use what their friends use. i use wicd, because i have for a long time and i want convienence from the command line. like netman users, i dont want to manually fiddle with wpa_supplicant configuration.

but i dont like it when i need graphics (or another open window, or hypothetically-- a reliable systray) to establish a network connection.

but this is about adopting netman. i will probably use it, the next time i have to recommend/install a graphical network manager for someone. its been a while, because im still waiting for a devuan source dvd like debian has (not to mention for release, not to mention for beta) before im regularly installing devuan for other people.

so edward, youre getting way ahead of yourself with that question. it will mean more when devuan itself is adopted-- yes, i know you can use netman independently, but probably devuan/refracta/aitors distro users will be the first-- i mean, most people will want an operating system well before they start needing a network manager.

ive been watching netman almost since the day you announced it, and ive been using devuan (and writing my own programming language) since the spring, and the only person i know (except online) using either of those is me.

not that theres anything wrong with asking now, of course. but it wont be a question that has a meaningful answer until 6 months, a year, 2 years from now.

its really too early to tell-- dont you think?

thank you for creating this wonderful resource and keeping our options open (without dbus even, as i understand it-- very nice.) when someone creates a friendly/or simple enough command line/or ncurses frontend, i will be more likely to use it myself. if you really want to speed up adoption, break wicd. no dont, that would be the redhat way-- we know youre nothing like them (and thank you.)

but im usually a slow adopter. there really are more of those than the rest.

i first used ubuntu in 2005-- it took me two more years to get rid of windows. id been exploring the landscape (and trying to use the software) for about 8 years already.

8 years? well, i was trying to replace my entire programming/file management/command line/browsing/graphical/hardware installing workflow. and i never did get dialup to work, so i had to start using real networking also-- something i didnt do before ubuntu.

im bet it will be a while before you read this, too. thats ok, it supports my point. thanks again-- like devuan, your project has fans! not all fans are users, and certainly not yet. cheers.
User avatar
fig1
 
Posts: 41
Joined: September 14th, 2015, 4:27 am

Re: adoption takes a while, when it happens

Postby nauved » November 20th, 2015, 3:26 pm

Adoption may happen more quickly than you think. Loved jaromil's recent comment to the list:

P.s back from blackhat EU here in Amsterdam
systemd is the inside joke :-D
and with it redhat is burning a lot of bridges in the enterprise sector...
They are going to loose the cloud race!

IIRC RedHat recently cut a deal with MS. So that might be the direction they're going to go.
May the FORK be with you!
nauved
 
Posts: 81
Joined: January 26th, 2015, 7:42 am

Re: adoption takes a while, when it happens

Postby nauved » November 20th, 2015, 5:02 pm

An update on the Devuan beta from irc:

<golinux> nextime: Progress on the beta?
<nextime> golinux: i'm working on the new amprolla
<golinux> nextime: And the systemd libs will be gone?
<nextime> golinux: : yep

Hooray!! Wonder if cups etc. will be fixed though . . .
May the FORK be with you!
nauved
 
Posts: 81
Joined: January 26th, 2015, 7:42 am

Re: adoption takes a while, when it happens

Postby cynwulf » November 20th, 2015, 10:23 pm

fig1 wrote:but how is devuan adoption going? well, its slow-going.

Why is adoption important? If the code is good, it will be used. If you like it, then use it.
cynwulf
 
Posts: 2548
Joined: April 26th, 2011, 2:46 pm

Re: adoption takes a while, when it happens

Postby fig1 » November 21st, 2015, 1:17 am

Winston Smith wrote:Why is adoption important? If the code is good, it will be used.


i cant speak for the author, but he implied in the title of his email that he wanted to know if his time was well-spent.

there are lots of good arguments that yes, of course his time was well spent. but in the terms he wants, i think its too early to even ask. (by now, at least one person on the ml has said as much.)



nauved wrote:Adoption may happen more quickly than you think.


i only meant its "slow-going" at the moment; the init-from-hell weve worked so hard to get away from spent at least half a decade as a minor curiosity, before snowballing into the full pita it became last year.

you would think pushing the avalanche back up the mountain would take longer, but i share your confidence that noticeable progress is right around the corner.

after all-- not only do i follow the mailing list and irc, i also run updates from apt.
User avatar
fig1
 
Posts: 41
Joined: September 14th, 2015, 4:27 am

Re: adoption takes a while, when it happens

Postby cynwulf » November 22nd, 2015, 2:57 pm

I searched and found the mailing list you are referring to http://news.gmane.org/gmane.linux.devua ... utoff=6185

It seems like it won't build and is not well documented.

In a general sense, "is my software being used or shall I give up?" comments do not instil much confidence. Which is why I said:
Winston Smith wrote:If the code is good, it will be used.
cynwulf
 
Posts: 2548
Joined: April 26th, 2011, 2:46 pm

Re: adoption takes a while, when it happens

Postby nauved » November 22nd, 2015, 3:22 pm

Winston Smith wrote:I searched and found the mailing list you are referring to http://news.gmane.org/gmane.linux.devua ... utoff=6185

It seems like it won't build and is not well documented.

In a general sense, "is my software being used or shall I give up?" comments do not instil much confidence. Which is why I said:
Winston Smith wrote:If the code is good, it will be used.

And the code will be good and do what's it's supposed to. The comment you quoted is irrelevant to anything but the author's need for validation for the work he did on one application. It has nothing to do with the quality of code of that application or confidence in the Devuan project.

BTW, posting this from a rock solid Devuan alpha2 install.
May the FORK be with you!
nauved
 
Posts: 81
Joined: January 26th, 2015, 7:42 am

Re: adoption takes a while, when it happens

Postby fig1 » November 22nd, 2015, 3:45 pm

Winston Smith wrote:It seems like it won't build and is not well documented.


why would it be well documented? he practically finished writing it yesterday. thats why its such a strange question for him to ask now-- good suggestion to document it better though.

fwiw, i have every confidence that when its packaged up (as people are volunteering, as debian packaging was never a treat) it will do what it says on the tin-- compilation included.

but you know, devuan itself is beta (or ever closer to it) and the documentation is still coming along, but then so is the thing being documented.

i dont think anyone is just sitting around-- im testing devuan constantly, both jessie and ascii and on 32 and 64 bit machines-- while writing (then documenting) two computer languages. not to mention constantly keeping up with devuan development.
User avatar
fig1
 
Posts: 41
Joined: September 14th, 2015, 4:27 am

Re: adoption takes a while, when it happens

Postby cynwulf » November 23rd, 2015, 9:14 am

nauved wrote:And the code will be good and do what's it's supposed to.

If he just gets on with it and stops looking for 'moral support' yes. What he did in fact was open the doors to criticism of the code and the fact that it won't build. Did you read the mailing list thread? Someone also mentioned documentation.

fig1 wrote:why would it be well documented? he practically finished writing it yesterday. thats why its such a strange question for him to ask now-- good suggestion to document it better though.

Documentation is important if you want others to build and use your program. It's best to start documenting as you go along rather than doing it retroactively.

fig1 wrote:fwiw, i have every confidence that when its packaged up (as people are volunteering, as debian packaging was never a treat) it will do what it says on the tin-- compilation included.

Packaging it won't actually resolve that. It should build and install with autotools before it's packaged. That's what he needs to work on and that's the best advice anyone can give him.

fig1 wrote:but you know, devuan itself is beta (or ever closer to it) and the documentation is still coming along, but then so is the thing being documented.

Isn't Devuan a collection of upstream software which just uses the upstream documentation? Surely, it's up to the developer to provide that, not Devuan.

Some of the comments seem petty however. He has built the frontend in object pascal, because that's the language he's familiar with and already someone questioned it. I don't see the language which the GUI is written in as that important for something like a network management utility.

I would also suggest he find a new name for the project as netman is the MS Windows network connection manager. Probably a coincidence, but a very unfortunate one.
cynwulf
 
Posts: 2548
Joined: April 26th, 2011, 2:46 pm

Re: adoption takes a while, when it happens

Postby hjkdadg » November 23rd, 2015, 10:16 am

nauved wrote:An update on the Devuan beta from irc:

<golinux> nextime: Progress on the beta?
<nextime> golinux: i'm working on the new amprolla
<golinux> nextime: And the systemd libs will be gone?
<nextime> golinux: : yep

Hooray!! Wonder if cups etc. will be fixed though . . .


Yep, close to beta since almost a year now ;)

Seriously, even the people who wrote the weekly newsletter are gone. git.devuan.org looks more or less abandoned (every few weeks something happens). They have problems to follow security updates, e.g. the one for dbus hasn't been integrated for months, though it should be trivial to rebase their changes on the current dbus package provided in Debian Jessie.

I expect it to be hard to have a release at all when the project looks like this, but they keep saying things like how they have 500 active developers, how they will do their own release next year (not jessie, but the one after). Sounds very dubious.
hjkdadg
 
Posts: 5
Joined: February 22nd, 2015, 10:44 am

Next

Return to Technical Talk

Who is online

Users browsing this forum: No registered users and 2 guests

x