OpenWrt Forum Archive

Topic: Improve the Wiki Table of Hardware?

The content of this topic has been archived between 12 Sep 2015 and 6 May 2018. Unfortunately there are posts – most likely complete pages – missing.

More cleanups and updates in Wired ports section.

zo0ok: Please update, TIA! smile

Wow, the Wired Ports really start looking organized, but still work to do.

...it had to be a German and a Swede to get this mess in order, didn't it, tmo26... wink

tmo26, there are 660 distinct device pages for our 907 rows (data entries).
That was worse than I thought.

I mean, the prospect of merging data entries... mostly each device is practically unique.

By the way, there are about 1200 image files in BB... but many devices have more than one file (typically factory+sysupgrade).

zo0ok wrote:

Wow, the Wired Ports really start looking organized, but still work to do.

...it had to be a German and a Swede to get this mess in order, didn't it, tmo26... wink

Yes, definetely! smile

zo0ok wrote:

I wonder how many device pages there are in the wiki, that are "orphanded", that is, without a link from the ToH. And how do we find those?

Today I stumbled upon a device with info only available in the oldwiki (mini something...; oldwiki page found via wikidevi. Nice to see that wikidevi is telling openwrt where the openwrt devicepage is located... ;-).

Another one not listed in toh, but devicepage available (ARV4525 / ARV4510). OK, very rudimentary, but still existing.
Found them via the manufacturers namespace http://wiki.openwrt.org/toh/arcadyan.
I hesitated to add them all to the toh, since image availability in CC / CC trunk seems unclear (only ARV4525 in trunk) and the overall available facts are very little.

zo0ok wrote:

Finally, if we could map "images" to devices that would be the absolute master to what devices are actually practically indentical and not. And perhaps our documentation should be based on images instead?

It would indeed be a big help to know which image for which device. The current status / devicepage / download is not always clear regarding this.

I'd say: Let CC15.05 final come, then we start collecting image urls, ok?
And maybe by this, the devicepages/devices-ratio can be improved, too.

Geeeeez, I had a long run through the wired ports column in supported + possible... maaaaany many updates!

zo0ok: I earned an update *sigh* smile

tmo26 wrote:

I'd say: Let CC15.05 final come, then we start collecting image urls, ok?
And maybe by this, the devicepages/devices-ratio can be improved, too.

Yes, I think something like that makes sense...
I updated. Back to 909 devices. Is it you that keep adding devices all the time? wink

zo0ok wrote:

I updated. Back to 909 devices. Is it you that keep adding devices all the time? wink

Rarely. Now and then. Last one being 4525-something, because there was an image available, but it wasn't listed.

Thanks for the update! smile

More cleanup of wireless ports on WIP page and several other small corrections.
I think we're slowly coming to an end.

zo0ok: Update please!

tmo26 wrote:

zo0ok: Update please!

Done!

I eliminated links to external device pages from the ToH. This brings me to gateworks...
I created: http://wiki.openwrt.org/toh/gateworks
(not following any template whatsoever)

In ToH:
http://wiki.openwrt.org/toh/start#gateworks

For Avila and Cambria, images appear to exist.
For Ventana and Laguna no images exist, unless the generic images are good.

In the supported column there is a link to http://trac.gateworks.com/
Hardly very helpful I think.

So
Avila/Cambia: can we just call them "supported" without the patches confusion?
Ventana/Laguna: either there are images (which images?) or they are possible/wip?

Ideas?

Set status in ToH:
Avila=12.09
Cambria=14.07
Laguna=14.07
Ventana=?

This is based on availability of images from those releases.

Did several more cleanups and additions and now need an update to see more clearly again. zo0ok, would you be so kind?-)
The changes are getting smaller, USB and Wired Ports are getting tidier.

What's your overall opinion of the current status, looking at the filters?

tmo26 wrote:

zo0ok, would you be so kind?-)

What's your overall opinion of the current status, looking at the filters?

It really looks good... that is my overall opinon. Of course, when it comes to UBS and Wired ports there is more work to do before we have 100% consistency. But that is when looking at the filters, and many devices are under unsupported/unknown/possible, and how much effort should we put into finding USB and Ethernet specifications for those?

When it comes to platform and wireless hardware there is probably also work that can be done. But it seems of quite limited value to the end users to improve those columns even more.

Also, the ToH will always be work in progress, so even if we make it perfect now, it will change in the future. 100% consistency is not necessarily worth it here.

I think we should start looking more into:
o status
  - is there an image?
  - is it working?
o target
  - where is the image?
o detail page
  - does it exist
  - does it contain information whether the device works as indicated in "status"
  - is there an image link

I am particularly unhappy with:
1) the ramips situation (where I contributed to extra mess, but where it looks very nice with 6 subtargets in CC)
2) brcm47xx (and the brcm-2.4 confusion, which I think should only be in the device pages, not in the ToH)
3) routers supported by "generic" image but this information is nowhere to be found

But (at least I) volunteer to do this job, and if it makes you happy to work more with the USB column I will not stop you wink

Well, I did update too.

(Last edited by zo0ok on 25 May 2015, 22:44)

I have spent a week or so mulling over this note, deciding whether to throw a rock into the pond or not. I am going to throw it.

(@zo0ok and @tmo26 and @all: Please note that I respect all the work that has gone to getting us where we are, and that I think there's a way to leverage it to make things even better.)

But I think we're missing an opportunity. As I read the tmo26's response at https://forum.openwrt.org/viewtopic.php … 7#p276557, I realized that I have had the wrong mental model about how things work today. If I understand correctly, this is the procedure for adding a new device to the ToH:

1) Create a new page that only holds the dataentry
2) Create a new detail page from the template at http://wiki.openwrt.org/meta/template_device. Fill it in with the information about your router: tech specs, installation and debricking procedure, pictures, etc. Use datatable's to get specs from the dataentry for the router
3) Create a new row in the ToH page, and fill in the columns for their new router. Make the first column of the ToH link to the new detail page. Manually fill in all the other columns with info about the router.

This seems confusing/problematic in several ways:

- The dataentry page (#1 above) exists solely to hold the dataentry. It serves no other purpose.
- The dataentry page is separated from the details page, so they may not stay in sync, and they'll certainly be hard to keep track of.
- We haven't stated (at least, I haven't seen) any rules for where (in the file system/wiki) it should be located. Even with guidance, they'll wind up all over the place. It's bad enough that people can put the details page anywhere they want, but a separate dataentry page will compound the problem.
- Information in the ToH needs to be manually entered - duplicated from the info in the dataentry. This violates the DRY (Don't Repeat Yourself) principle - since they will inevitably get out of sync.
- zo0ok runs a manual process that produces the very cool variant of the ToH, but it isn't available to people using the site.

I see an alternate way to do this. I realize what I am proposing isn't how things work today, but please hear me out. I have two observations/desires:

1) I really would like the dataentry to be part of the details page. This keeps the router's details together with the dataentry, so they can be edited in a single operation. And the dataentry formatting on the detail page is both functional and beautiful (see, for example, http://wiki.openwrt.org/toh/d-link/dir-505 )

2) It would be great to build the ToH page from the information contained in each dataentry. This means no manual duplication/errors, and I think it would be simpler for people to update/add to the ToH. It would be especially cool if we could add a some CSS/formatting to zo0ok's automatically generated page to make it the default ToH.

Is this possible? Am I way off track here? Thanks.

(Last edited by richbhanover on 26 May 2015, 21:56)

richbhanover wrote:

I realized that I have had the wrong mental model about how things work today. If I understand correctly, this is the procedure for adding a new device to the ToH:

1) Create a new page that only holds the dataentry
2) Create a new detail page from the template at http://wiki.openwrt.org/meta/template_device. Fill it in with the information about your router: tech specs, installation and debricking procedure, pictures, etc. Use datatable's to get specs from the dataentry for the router

So far so good. (except the tech specs, they go into the dataentry page)

3) Create a new row in the ToH page, and fill in the columns for their new router. Make the first column of the ToH link to the new detail page. Manually fill in all the other columns with info about the router.

No. It would be mad to first create the dataentry pages and then manually change the toh.

You still havn't fully realized how the new toh will bre created... smile

To say it clearly: The new toh will be created automagically (...out of the dataentry pages)!

You create a new dataentry page and it will automagically show up in the toh. No further manual interaction necessary.
That's the whole sense of this dataentry-page effort: Have the technical specs of a device in one and only one place, and use this data flexibly all over the wiki (first of all in the devicepages).

- The dataentry page (#1 above) exists solely to hold the dataentry. It serves no other purpose.

No, it is the datasource for the automagically created toh. That's it's purpose.

- The dataentry page is separated from the details page, so they may not stay in sync, and they'll certainly be hard to keep track of.

No, dataentry and detail page always stay in sync since the detail page does not contain any actual data. The detail page instead uses data that is contained in the dataentry page.

In the sourcecode of my DIR-505-example in the playground (you overwrote it) you could have seen how this is accomplished on the detail page. It's dead simple.

- We haven't stated (at least, I haven't seen) any rules for where (in the file system/wiki) it should be located. Even with guidance, they'll wind up all over the place. It's bad enough that people can put the details page anywhere they want, but a separate dataentry page will compound the problem.

You can put the dataentry page basically anywhere you like.
I have already proposed to put all the dataentry pages in a separate namespace, at least for tryout purposes (easier maintenance).

- Information in the ToH needs to be manually entered - duplicated from the info in the dataentry. This violates the DRY (Don't Repeat Yourself) principle - since they will inevitably get out of sync.

No, see above.

- zo0ok runs a manual process that produces the very cool variant of the ToH, but it isn't available to people using the site.

Full ACK, I like it too!

1) I really would like the dataentry to be part of the details page. This keeps the router's details together with the dataentry, so they can be edited in a single operation. And the dataentry formatting on the detail page is both functional and beautiful (see, for example, http://wiki.openwrt.org/toh/d-link/dir-505 )

Doesn't work, for reasons already explained.
http://wiki.openwrt.org/toh/dataentry_t … _dataentry

2) It would be great to build the ToH page from the information contained in each dataentry. This means no manual duplication/errors, and I think it would be simpler for people to update/add to the ToH. It would be especially cool if we could add a some CSS/formatting

That's the way this whole thing is supposed to work in the future.
And yes, usage of CSS is possible.

to zo0ok's automatically generated page to make it the default ToH.

That would be an alternative solution. We can run it in parallel.


Is this possible? Am I way off track here? Thanks.

You are off track, meaning: You seem to have missed some postings and hence havn't understood the working principle of dataentries and "new toh". Hope I have brought you back on track now ;-)

(Last edited by tmo26 on 26 May 2015, 22:48)

zo0ok wrote:

100% consistency is not necessarily worth it here.

Fully agree. And 100% isn't desired / required.
I put together a crude statistic, where I assumed 10% error allowance in relation to the whole number of devices for certain columns (I posted about this some days ago, with % and threshhold etc).
My goal is to have the information as complete as possible - at least for the supported devices and if the information needed is acquirable with reasonable effort. Tonight I did some more updates (yeah, you guess it: USB and Wired ports again smile). For some devices it was easy, for some very hard.

If you would update once more, I would go through the devices a last time and then declare it for "best as we can get it now". And if you look at the data: We're pretty damn good.

I think we should start looking more into:
o status
  - is there an image?
  - is it working?
o target
  - where is the image?
o detail page
  - does it exist
  - does it contain information whether the device works as indicated in "status"
  - is there an image link

Yes please! All of that! smile
How / where do we manually collect the fw image urls?
Is there a way that this can be done at least partially automagically?

Can the toolmaker forge some special tools out of "Schwedenstahl"?
(for specs of "Schwedenstahl" see http://www.smt.sandvik.com/en/products/ … ik-12c27/) ?-)
(forgive me the allusion to your location and your programming talet *g*)

2) brcm47xx (and the brcm-2.4 confusion, which I think should only be in the device pages, not in the ToH)

Then clean it up, it's not that much. Now is the time to do it smile

But (at least I) volunteer to do this job, and if it makes you happy to work more with the USB column I will not stop you wink

Do you know Monk?
It needs to be 100. Not 99, not 101.
Different to Monk, there are times when I give up. ;-)

Well, I did update too.

As said above: Once more, please!
Not many more expected. I'm getting really tired and want to come to an end.

Update complete: https://dl.dropboxusercontent.com/u/906 … index.html

@richbhanover and @tmp, I summarize as I understand it. If we are lucky, you both agree with my understanding, and we are all in sync. Otherwise it does not hurt to discuss this more.

The soon to be "Table of Hardware" and "Device Pages":
- The ToH as we no it will go away
- Each row in the Table of Hardware will have its own Data Entry Page
  (with the same, but also more, information than todays ToH)
- The future ToH will be dynamic and entirely based on the Data Entries
- The Device pages... well at least they will not contain the information that is already in the Data Entry because they can "import" values from (one or more) Data Entry and display it.

So I think we have two important tracks to proceed on:
1) status/image/target-work (and probably it depends on that we get started with Data Entries, since that is where the image link goes)
2) for some (complicated) devices, work with the DataEntry-DynamicTableOfHardware-DevicePage model so that we dont migrate data and then regret the model.

Have I got it quite right?
@richbhanover: does it make sense? (btw, I like throwing stones in ponds too)

Edit: one more little reality check here...
I, @zo0ok, can write scripts to help automate the generation of the Data Entries
It is @tmo who understands that the above will work as described and intended in the wiki. I dont know that, and I am not a wiki expert.

(Last edited by zo0ok on 27 May 2015, 08:15)

zo0ok: Thanx for the update!


tmo26 wrote:

If you would update once more, I would go through the devices a last time and then declare it for "best as we can get it now". And if you look at the data: We're pretty damn good.

Went through USB + Wired ports again, did some final cleanups.
After this, I hereby declare the current ToH for "best as we can make it now". At least from my point of view.

zo0ok: final update, please!

Champaign for all who contributed! ;-)

Sorry, posts 351 to 350 are missing from our archive.