Dear Bob, thank you very much for such a brief, detailed and helpful reply! It has cleared most of my questions, but there are a few more:
1) Although I fully agree & support the five from six conditions list (especially those which are about proven absence of closed source blobs) there are some doubts about this one: > > * The target must be flashable from stock firmware without any > elaborate hacks. > In my humble opinion, if a device successfully meets five other conditions, then it doesnt matter if a user should do some tricks to install LibreCMC - as long as those tricks are well researched and documented. Also, even if its impossible to flash a target straight from a stock firmware - if this device meets all other five conditions, there is no much problem: many boards could be flashed directly by dirt cheap $1.5 USB to TTL adapters (shipping from China included) 2) Recently I learned about LibreCMC project, would like to buy a new router and get involved! But here is a problem: the only relatively modern router that is listed right now on "Supported Hardware" page and has 2.4GHz + 5 GHz support, is a Netgear WNDR3800 - which is overpriced and almost sold out everywhere. So I researched about all the routers listed in config file, and it looks like in this list there only two routers which satisfy to "affordable price"+"at least some availability"+"2.4/5 GHz" requirements : TP Link TL-WDR4300 TP Link TL-WR2543 Please tell, if any of those routers is already supported by LibreCMC (but it could be hard to install and thats why not listed officially) ? Or are there any potential supported 2.4GHz/5GHz LibreCMC candidates which haven't been added yet ? On 10/09/2015, Bob Call <b...@librecmc.org> wrote: > On Wed, 2015-09-09 at 22:48 +0300, Vladimir wrote: >> At supported hardware list page, there are only ten router models >> right now: >> >> 1) Buffalo WZR-HP-G300NH ; 2) Buffalo WHR-HP-G300NH ; 3) Netgear >> WNR2200 ; 4) Netgear WNDR3800 ; 5) TP-Link TL-MR3020 v1 ; 6) TP-Link >> TL-WR741ND ; 7) TP-Link TL-WR841ND ; 8) TP-Link TL-WR1043ND ; 9) >> ThinkPenguin TPE-NWIFIROUTER2 ; 10) Qi-Hardware Ben Nanonote >> >> However, at your librecmc-config file there are a lot more targets >> available, such as TP-Link TL-WR2543 . And at your >> snapshots/v1.3.1/ar71xx/generic/ directory there are many successful >> builds available, such as for: Buffalo WLAE-AG300N , Cisco-Linksys >> WRT400N , Buffalo WZR-450HP2 > > While these targets "build", they are not officially supported by the > libreCMC project. In order to be classified as being officially > supported, the following conditions need to be met : > > * The target can't require any non-free blobs. > > * The target needs to be formally tested. > > * u-boot or another bootloader needs to be flashed and verified that > it does not contain any blobs. Hint: some switches may need blobs that > are loaded by u-boot. > * Most, if not all, versions of a device line must be supported (to > avoid confusion). > > * The target must work in a reliable fashion. Hint: most embedded > devices don't work well or the quality is poor. > > Yes, I should have added a subtarget for unsupported platforms, but > I did not. > >> Please update your "supported hardware" page: if more people will >> know >> that their devices are compatible with LibreCMC (even if its hard to >> install it on some of devices without specia[l] tools like flashers) >> then >> LibreCMC will receive more users and gain popularity! >> > > Just because a device appears on a hardware list, supported or > otherwise, it will not increase the popularity of the project. The goal > of libreCMC is to provide a fully free embedded OS (GNU/Linux or > otherwise), not to be part of a larger "popularity" contest. If > libreCMC becomes popular, that is great, but the project's goal does > not include popularity. > > Right now, the last thing the libreCMC project needs is to > [artificially] grow out its user base. The focus of the project, right > now, is to clean things up and make the OS/distro [1] smaller. libreCMC > will gain popularity over time because it is fully free and because of > its technological merits, not the list of shitty hardware that it > [won't] support[s]. > > Sure, it would be nice to add a boat load of supported hardware and > software, but what would the cost be? Most embedded devices that > upstream supports don't work well (in addition to having freedom > issues). By supporting hardware without extensive testing, this sets > the libreCMC project up for being seen as having poor quality standards > or being "buggy". People will blame their software well before they > blame the hardware they are using (even with proof). The same goes for > adding features, if it does not work well, it won't be added or > supported. > > >> Also, its better to set up this "supported hardware" page in table >> style, similar to what coreboot project wiki has. That will allow to >> provide much more important information, such as supported revisions >> of boards of some router models. (each router model could have >> several board revisions, and it could be that not all of them are >> compatible) > > Our current model of vendor lists with supported device lines is more > optimal. libreCMC only supports device lines where all targets are > supported (there may be an exception to this rule, but it is made > clear). Yes, the documentation on the libreCMC wiki needs to be cleaned > up, but I don't agree with the proposed model. > > [1] libreCMC is exploring other OS models that don't use the linux[ > -libre] kernel and would no longer be classified as a GNU/Linux[-libre] > "distro". > > -- > Robert Call (Bob) > b...@librecmc.org > https://librecmc.org > _______________________________________________ > Librecmc-dev mailing list > Librecmc-dev@lists.librecmc.org > https://librecmc.org/mailman/listinfo/librecmc-dev > _______________________________________________ Librecmc-dev mailing list Librecmc-dev@lists.librecmc.org https://librecmc.org/mailman/listinfo/librecmc-dev
