RDF Packaging and Distribution Poll

[Capn Trips]

In response to the running thread here, I've created this poll to help WagonMaster detemine how to go with the packaging and distribution of the RDF Zip file.

[ElizabethD]

While being mindful of the distinction from the other thread,

Be sure to distinguish distribution (my issue, as RDF maintainer) and installation (your issue, as a user). This discussion is about distribution, not installation. If you install both aggregated RDF zip files (unextended and extended) into the same directory (which, of course, you should), the operation of the utilities won't change a bit.

I'd still prefer ONE zip file for downloading.

[unclemiltie]

As an extender writer, i would prefer that the base RDF's do not include the RDF's for the extenders. Every extender distribution that I have seen has included the RDF's.

this gives the extender writer the ability to tie the extender to the RDF that ships and thus less possibility of errors.

(so I voted to split, but I'd prefer that tehre not be a central repository for extender RDF's)

[vickyg2003]

I thought unclemiltie wasn't going to like this!

As an extender USER and Writer, I like the idea of RDF's that have been SUBMITTED by the extender writer be included in the main RDF zip file. I remember that it was important to delete my folder and do a clean install with a new release of RDFs. When I do that, I'd be missing the RDF's for all my remotes. I might as well skip the download all together. I use the 7800, 6131, 6011, 8811, 8900, Comcast 1067a, Atlas 5. Assuming WagonMaster is going to be issuing RDF's more frequently this becomes drugery.

I also like the idea that with any changes my USERS would be getting updated RDF's without having to do a download everytime there are updates that would make my RDF's incompatible with IR or RM.

[WagonMaster]

Setting up this poll was a great idea! My thanks to Capn Trips for doing so.

As I stated in that other thread:

I'm going to refrain from voting for the moment, however, because I think my vote depends on the outcome of the discussion about whether we can unambiguously detect an extended RDF. Once that issue seems decided, I'll add my vote to the mix.

Thanks again, Capn!

Bill

[WagonMaster]

I remember that it was important to delete my folder and do a clean install with a new release of RDFs. When I do that, I'd be missing the RDF's for all my remotes.

Vicky, it might be worthwhile for you to investigate "symbolic links" as a way to circumvent this. Unix (Linux, et al) has had that feature for ages and I seem to recall that Windows finally added a poor-man's variant of this some time back (Win2000/WinXP era). Check the Wikipedia article for more info.

If that doesn't work well enough, why not set up a simple batch file to copy all your extenders from 1 spot to another?

Bill

[ElizabethD]

Here's the deal.
The extender writer includes RDF file in the distribution.
Years later the extender writer is fishing instead of doing JP1.
Something incorrect is discovered in the RDF file, or a global change is made to all files in one place.
Nils would correct, put it into the main zip and people were told to always check for a most recent RDF and ignore the RDF inside the extender package.

If RDFs are with the extender then they would need to be maintained on a one-by-one basis in files that somebody else owns.

[mr_d_p_gumby]

Years later the extender writer is fishing instead of doing JP1.

So that's what happened to them!

Nils would correct, put it into the main zip and people were told to always check for a most recent RDF and ignore the RDF inside the extender package.

I don't know for sure about some of the most recently created extenders, but ALL of the older ones basically used the same slightly-edited readme file that specifically told the user that they should always download the most recent RDF file in preference to the one that was included with the extender.

If RDFs are with the extender then they would need to be maintained on a one-by-one basis in files that somebody else owns.

That's the problem. No one is going to do that to someone else's extender distribution because then they assume some responsibility for the entire extender, not just the updated RDF.

[unclemiltie]

Mike, Liz and vicky, I can see your point.

however, I'm still leaning toward keeping extender code and extender RDF's tied together. I can say from experience that I've had to circumvent the RDF that was distributed in the master distribution with a new one when I released a new extender version. Versioning has helped since IR 8 came out, it shows the extender version (it is embedded in the extender code) and the RDF version (it is embedded in the title)

But no matter how we do it, there are always going to be mixups. The only question is which method makes less of them!

[wnewell]

Frankly I don't care how you do it or how many are in it, as long as the ones that are work. I say make it easy on yourself. If they are already in it, leave them there., or not.

[Capn Trips]

I agree with Liz. In over 4 years of using extenders with various remotes, the maintenance of the "master" RDF's seems to be at least slightly (if not greatly) more reliable than similar maintenance of individual extender files. Of course more recenty developed extenders are logically more likely to have their authors still here on the site and updating their individual extender packages, but I would wager that most of the older (pre-JP1.X) extenders have not had their distribution packages updated in years. Therefore, to WHOM should fall the task of bringing those older extender RDF's into compliance with evolving RDF specs? If I'm running the RS 15-1994 with extender 5, WHO will update that RDF to the new RDF spec? NOBODY - unless the RDF is inculded in some sort of "master" distribution.

I believe that this RDF maintenance imperative should take precedence over the potential additional coordination required for currently "active" extender authors to reconcile any changes to "their" RDFs necessitated by either extender evolution (initiated by the extender "owner/author") OR by a change to the RDF spec (initiated by the RDF file "maintainer".

[mdavej]

After reading all the arguments, I think a single distribution for all is best. It adds somewhat to a newbie's confusion when making upgrades (which remote to pick), but improves the chances of better maintenance as other's have pointed out.

In cases where there are multiple versions of extenders, each with different capabilities, like the RCU810, what gets included in the master? All versions or only the latest? How do we know if errors are introduced?