Who is our Audience? Who uses (will use) LEDE?

As we kick off the LEDE Project, we're having good "startup discussions" about how we want to organize things.

Rather than let the discussion spread across lots of unrelated topics, I want to try to consolidate them into one place here in "Site Feedback".

This is the first of several a few questions about how we want to construct and evolve the site.

Who is our audience?

  • What level of expertise do we expect them to have to use the site
  • What are they trying to do?
    • Just try LEDE?
    • Build a solid router for home (and family) use?
    • Testbed for learning about networking/VPN/firewalls/etc?
    • Build LEDE with their own package set?
    • Create new packages?
  • How often do they expect to install LEDE?
    • Install it once and leave it alone?
    • Install it a couple times over 3-6 months then let it sit?
    • Create many builds over time?

The answers to these questions will help us prioritize the effort on pages of the site.

I started using openwrt on a wrt54gs with exparimental WR trunk builds.
I've updated my router every 3-6 months (when I bother to remember) since then from trunk/master.
I've made a couple appliances over the years for random people on IRC. But until last week I haven't needed to add any hard packages to my systems.

You can create polls for such purposes:

  • Install it once and leave it alone?
  • Install it a couple times over 3-6 months then let it sit?
  • Create many builds over time?

0 voters

Actually, I was hoping that the leaders of the site would weigh in before sending it out for a poll. Thanks.

currently, the userbase is shifted more towards the high-end power-user Windows user, or average Linux user.

Most people I know tend to make trunk builds because OpenWRT development times were kinda slow and most cool stuff (packages) was always in trunk.

OpenWRT was/is mostly sought after (again in people I know and see/interact with around the net) because (in no particular order):

  • it is opensource therefore trustworthy for security/privacy

  • it receives updates and keeps secure even very old devices that would be EoL since like 4 years but work fine. (may be not 100% fast on this but anyway it does a better job than the manufacturer)

  • it allows to have pro-router feautures like true badass VPN servers, support for ALL Dynamic DNS (not just 2-3, curse you), VLAN with a pro-switch-grade GUI too, and decent QoS without shelling out 300$ for a gold-plated octa-core warp-9.99 wifi router.

  • its secondary (non-networking) features are actually reliable and work EXACTLY the same on all openWRT routers (media sharing features, 3G/4G dongle support, various servers) so once you learn how to configure them ONCE you know how to configure them any other router. That's a BIG thing if you are into using LEDE in a techsupport shop for consumers or if you are the "techie friend" of many people. Most smart router manufacturers lock you in with consistent firmwares (like Asus, actually not even bad, they are probably the only one that is actually designed to support Entware packages, a fork of OpenWRT package feeds), some are just plain inconsistent even within their own products.

  • because they tried to do something that isn't basic stuff with dd-wrt (like say VLAN, or trying to host a ftp server, just try yourself) and noticed the horrible mess that dd-wrt is beyond basic usage and giving you good wifi performance on broadcom, and how total garbage is dd-wrt wiki (manages to be worse than OpenWRT's, personal experience) see for example here and here

AFAIK most of the mesh wifi thing is german-only that happens in a few downstream projects, I didn't usually see people talking of mesh wifi and OpenWRT, they are all using some downstream fork with a funky name like freifunk or whatever (there is a list of them in the LEDE wiki main page as Endorsements).

(little OT, freifunkers have built a webapplication that allows them to make custom firmware images using the Image Builder, that would be very convenient to have for us too.
here the site http://meshkit.freifunk.net/ here the source https://github.com/weimarnetz/meshkit setting it up is a bit beyond my abilities though)

please provide the data that you are basing this on. to me it sounds like a gut feel that you have and there is no evidence to support this theory

[quote="blogic, post:6, topic:185, full:true"]please provide the data that you are basing this on. to me it sounds like a gut feel that you have and there is no evidence to support this theory[/quote]Well, I work as techsupport since a few years and I've been active in english and italian technology forums for like 15 years so I think I can claim to know the broad categories of users. I extrapolated on that.

  • Most people that actually know what a router or NAS is are windows power-users, people that work in IT or have a hobby that requires them to know more than average.
    Most people using devboards or doing any other kind of project with a generic device that does not have a clear design goal in mind (like a raspberryPi or also Arduinos) are also in this category. They are mostly windows users that may have learned a bit of linux for their project (Arduino is not linux I know), usually none of these has a linux PC, they work on linux devices from windows.
    .
    These people are interested in better/reliable features, more granular/advanced configuration, security updates (all things you don't usually find in stock firmwares).
    Currently, as more and more router manufacturers wake up to this (Asus is one, at least for the secondary packages in routers, Synology or QNAP are huge in the NAS market due to their similar approach to Android, they provide apps for the NAS to extend functionality) less people in this category will be interested in geting a OpenWRT/LEDE device.
    .
    These users can be divided in "tinkerers" or "technicians" below.
    .

    • "Tinkerers" accept relatively complex procedures to hack their device to accept more functionaliity (like for Optware/Entware) as they like the "hacking" part.
      .
    • "Technicians" don't like hacking up devices so the faster and easier it is to get to the goal, the better. People wanting better performance for their NAS, gamers that are sharing the network with others and need QoS, and so on.
      .
  • Most linux users I see around/know/interact-with fall in the same categories, and for a linux user there are also other reasons on top to like/try LEDE, the fact that it is opensource therefore more trustworthy than stock firmwares.
    This reason to use LEDE is still as strong as ever.

  • As for macOS users (Apple), it's the same as with windows power users.
    macOS is more developer-friendly than Windows while also much more stable and usable as-is than linux for a simple end user, so it attracts both developers (people interested in LEDE) and end users (people that don't know what a firmware is, and don't even care)

While I can't really pull numbers to show who of these is the more important, these are the most likely LEDE targets, and I think it's kind of obvious.

Most people outside of technology-savyy circles or forums just go and buy "a modem" (that's how they call "a wifi modem-router") of a brand they trust or just a random cheap one and they use it as a dumb tool to reach the Internet until it breaks.
All those that don't have a NAS don't even need gigabit ethernet or ethernet at all so in many many many homes you still find ancient hardware with 100mbit and wifi g or even the odd non-wifi modem-router or ethernet modem connected to the only PC in that house.

as i expected, personal observation with no real facts to back it up. you do realize that when you proclaim your perceived realities, that people will consider these to be accurate which they simply are not.

so the discussion of who the target audience is has been going on for a long time. originally mbm and groz wanted owrt to be a one stop shop for all your embedded needs. then over time owrt gained a lot of commercial attention and industry players started to want to have a say. as owrt/lede is deployed on many more vendor devices, one could argue that this is the biggest audience simply due to the quantity deployed. the next group would argue that home users replacing their stock FW are the main audience. another group would argue its targeted at developers as an enabler tool for things like mesh networking. there is no simple way to group these that i can tell. as we have various channels for contributing and participating, not everyone will be on all of them due to personal preferences.

the question that should be solved i guess is if this solely a forum for users or industry users or devs .... or everyone.

I feel the tension in this thread. Please all, let's adhere to LEDE Rule #12...

I started this topic to address this: The answers to these questions will help us prioritize the effort on pages of the site.

@blogic - where do you think we should focus our initial documentation efforts? Is there something we need to address first? (Obviously, we want to document everything in the fullness of time...) But where should we start? Thanks.

Imho the initial documentation efforts should cover the basic usage and configuration and the process of building LEDE oneself.

Data => Information => Decisions

We have the tools and can capture the data to get a real answer to this question, but need to capture the right data.

First, I suggest adding a survey to the sign up process for new users and collect information on their current hardware and expected use. As we already capture their email, we could periodically (annually or prior to a new release) remind users to update their hardware profile, inform them of new releases and request their help with other periodic queries that can be used to improve the product (ie your device uses this driver, how do you like it?).

Second, when a user downloads product from trunk, and eventually the standard releases, we can track these downloads by IP address and date. Tracking by IP will allow us to develop a profile for an anonymous user of what product they have and additional functionality they want (packages). Tracking by date helps us understand the frequency they update their devices. (We make some assumptions that they use what they download, but it's better than nothing)

Data can be aggregated by device and package to build statistics on the most popular devices and packages. We will be able to understand how many users are moving from factory to LEDE as opposed to upgrading (new vs established users) and many more aspects of the products usage.

This data will help not only to understand who the actual user is, but how LEDE developers can focus their time, what devices can be sunset and prioritizing the wiki effort.

Data analytics will answer the question and guide LEDE to better focus on product and support. I am sure some will consider this a bit "big brother", but the reality is this is good business and does not need to compromise the security of any user.

Structuring this data collection now is important. I will suggest that it is critical for this to be in place prior to any production release, so that we can accurately track all the downloads. Currently we have about 100 users who have no choice but to use trunk. The first release will hopefully radically increase this number and will be the best opportunity to answer the question of Who and Why?.

Site analytics should be able to collect all the data, however I do not expect that it will offer the ability to analyze this in real time. The data will most likely need to be "warehoused" and periodically compiled offline.

1 Like

Uhm, I think there is a misunderstanding.

I listed all possible users for LEDE, and the only ones I dropped from the list are people that don't know what a router/NAS is as they obviously won't be interested in LEDE.

My list is aimed mostly at making sure that none of what I know are potential users is ignored, not at focusing efforts on some perceived "best" group among them. I didn't state any such "more worthy" group.

If LEDE was a company that has to stay profitable, this is what I'd do too. It is right, correct, effective.

But LEDE isn't a company, most of its structure is (like most opensource projects) is based off unpaid voluntary contributions.

Even if you had data you would still have 0 authority to stop people from contrinbuting in whatever they feel like or telling them "go make X", so imho doing anything more serious of what there is already is a waste of resources.

I think the best approach is to keep in mind all parties, and do first the things that would be most useful for all.
What said jow for example (documentation about base system and building LEDE from source/with Image Builder) is most useful for everyone I listed, therefore should have priority.

A good example for IMHO useful statistics: Dokuwiki's "Popularity Feedback"

anon_id, Version info, some statistics: Who is using dokuwiki? Big site with thousands of big pages, or small blog with only a hundred small pages? It is not a bad idea to know the customers a little bit. Make it easy (click -> send), make it anonymous and optional, and present the results to the users (to satisfy their curiousity and to reward them for their efforts of sending the data)

Installed plugins: From the installed plugins, the popularity graphs on https://www.dokuwiki.org/plugins?pluginsort=^c#extension__table are created.

php extensions: Knowledge of the php extensions helps in solving issues

It will certainly need some effort to implement this Feedback, and even more effort to forge some useful statistics out of this feedback, but datacollection should start as early as possible. In the forum, the wiki, the downloads. Let's see what pictures we can draw from this information.

BTW: Who is our audience is answered here: https://www.lede-project.org/

What is LEDE’s target audience?

As a source based Linux distribution for wireless routers, LEDE is primarily
> intended for networking enthusiasts, wireless communities, embedded Linux
> developers and advanced users.

LEDE aims for versatility and ease of customization but does not attempt to
be a fully integrated alternative to OEM firmware, it rather is a set of
building blocks to implement your own projects.

It was me who wrote the text quoted by @tmomas - I tried to make it match the current reality of LEDE / OpenWrt which does not mean that I want to exclude any kind of audience.

My thought was more that it makes to sense to work from the bottom up, means cover the development and cli level configuration first, build a solid base of power users, then start making the project approachable for entry level users.

Reading the links for the Popularity Feedback plugin, it appears to me that what this is doing is sending our Wiki technical info (Template, Plugins, Webservers, OS, PHP version, Languages) to Douwiki for their usage stats.

I think it's the AWStats tools that get us LEDE site data. There is a thread here.

There are some other DocuWiki Plugins that will work on the WIKI, not sure if they will just target the wiki or the site. They do not indicate downloads in their descriptions.
https://www.dokuwiki.org/plugin:statistics
https://www.dokuwiki.org/plugin:statdisplay

When visiting LEDE forum I am often looking for wikis and knowledge on how to do things. I am also seeking out other people's recommendations and experiences to help me make good decisions in purchasing hardware and what/how to configure things. I also look for industry related news and discussions on things that affect the LEDE community.

I am interested in LEDE because it is orienting itself toward reliability/dependability. This is what I most want. I am primarily interested in building a solid router for home/family use. I want the best reliability possible and the ability to make use of https login to LuCi along with dnsCrypt installation and configuration on the router. I appreciate being able to make setting/configuration changes (e.g., being able to change public key strength for LuCi https login). I also like LEDE/OpenWRT because of the packages provided (e.g., https-setup and dnscrypt) in the LEDE/OpenWRT repos.

I do want to "learn" more about networking and firewalls but I see myself as doing this by getting the router built for my home/family use.

I like to install a Stable version of the software and then leave it. When a new Stable version of the software is released, I will update to it. I appreciate being able to "update" software, but usually when there are large gaps between stable releases I will backup (or take note of) of my current settings/configuration, download the new version, wipe/reset my current setup and then install the new version. In this way it serves as a "pristine" fresh install. But in-between Stable releases I like not having to update or mess with anything. Although if there is an important security update in-between Stable releases, I will update to incorporate this.

Why would a company have more use for that data then an end user. As end user this same data will answer the same questions. A long as the data is understandable it will be usefull.

As data is a source it is also nice to have that source open. You don't know what i want to know.
And what i know might help you to do other things that you didn't even know about.

Sure we should not waste resources, but you can also see it as an investment in the future, as you don't know what you don't know, But if somone else can figure it out ten you don't even have to think about it. So then it has saved resources after all.

That's a risk. but data is knowledge and knowledge is power.

People may not be able to contribute in your field directly, but to another field where other people are who in their turn may contribute back to your field or someone elses.

Bazaar style