[Ovmsdev] Default values and fields...
mark at webb-johnson.net
Wed May 23 09:49:25 HKT 2018
While I agree in general with the points you make, there are limitations on both the platform and the community.
Firstly, we don’t currently target this as an end-user device. We clearly say "Designed for technically experienced OVMS v2 users and v3 developers”. Main reason for this is to limit support calls, until firmware gets more mature, but also because it is relatively easy to make a developer kit when compared to an end-user product. We simply don’t have paid commercial support; everybody here are unpaid volunteers.
We could dumb-down the config setting for timezone, by not supporting DST auto-switch. A simple -12 .. 0 .. +12 dropbox selection. That is what a lot of consumer devices do. But, that just seems wrong if we have the capability. Does your HP 100LX handle DST auto-switch at all?
We can’t add nice country/city selection because of this:
$ du -sh /usr/share/zoneinfo
That is from a Linux box. We don’t have 5.5MB of flash to store all the timezone settings. Maybe we could do some api in the cloud to look-up, but that is quite some work.
There are the iana rules files, which are smaller at 1.1MB, but still too big. They include all the historical info that we don’t need. I guess we could write some sort of parser and library to get it, but not at all easy.
So, the options are:
Dumb it down, and drop DST switching
Leave it as it is
Someone take it on as a sub-project to make it easier (just a simple country/city -> posix string mapping)
P.S. As to the politics of timezones, I live in ‘HKT-8’ (POSIX string). No DST (since the 70’s). I have never really understood how DST can still exist in the modern age, and how the rules can be so complex. The cost of lost productivity must be horrendous.
> On 23 May 2018, at 7:36 AM, Greg D. <gregd2350 at gmail.com> wrote:
> Hi Steve, all,
> This totally highlights the problem. No offense intended to the developers (myself included), but we've not structured the product very well for "normal" folks who are somewhat technical, but have no inside knowledge of the product, prior versions, Linux, ESP-IDF, etc.
> How was I to know what to put in the Timezone field? Bear with me as I walk through my troubled mental process... And please take this as me trying to be helpful.
> // Begin Rant
> So, I've seen configuration screens a bazillion times asking for timezone. If there's not a pull-down, usually I would put in -7 for the US Pacific Daylight Time that I'm in right now, or -8 if on Standard Time, and make a mental note to change it next fall. Not a big deal; there are a dozen "clocks" here at the house that need semi-yearly care. What's one more? Sometimes the config would be -8 with a checkbox for enabling DST. More recent product config screens sometimes use cities, with Los Angeles being our "local" example. That's been the set, so glancing at (but not truly understanding the implications of) the fine print on the configuration screen, I simply put in my -7 and moved on. It's been that way for weeks, and I never noticed anything wrong until I saw the "GMT" reference on the Time at Boot status that I recently added. Should have been "PDT". Something's wrong. Playing the ever-so-observant beta tester, I dig in...
> So, back to the config screen I went. First, where was it? I was doing a firmware update, but it's not there, and there's no General config tab. Nothing labeled "Time". Finally found it under Vehicle. Really? I see my -7 still sitting there, so it wasn't lost. Finally I re-read the comment about a web link to "Example Timezone Strings", and think, huh, I guess there must be some other way to say "-7". I click on the link, and scan down the table. It says USA & Canada, Pacific Time, and PST8PDT. Odd, but ok. Copy & paste. Hit Save. Go back to the status page, and see PDT on the aforementioned boot status line. Also see the upgrade was done yesterday afternoon, and realize that was 2am GMT, so I guess it checks out. Annoying, but DONE.
> Now you tell me I'm not done, and that I really needed to click on the second link to the man page on glibc? And understand it? Then research what the rules for DST computation are in my country, state, and so forth, to craft some sort of command lingo for the TZ engine?
> Reality check here. My HP 100LX palmtop, manufactured in 1992, does a better job at handling this. We've taken a simple task of entering a small, well known signed integer, and turned it into an exercise in POSIX programming and international politics. This for a parameter that has little, if any, consequence on the daily operation of the product as I currently use it.
> // End Rant
> With a grateful acknowledgment to Michael for his awesome work on the Webserver UI, I think we should take a pass through the configuration steps with an eye towards making this as turn-key as possible. I understand that he's made some significant changes to the UI since my module arrived (back on .003 code), so the improvements in initial config have not been reflected in what I am running now (since the config has already been set). As I noted in my previous email, I will blow my module back to factory defaults and start over, looking for places where we might improve things.
> I know that the V3 OVMS module is a VAST improvement over the prior versions, even without having experienced them myself. But to reach a broader audience, we need to continually look to improve the setup / configuration experience, especially as we add new and even more powerful features.
> Stephen Casner wrote:
>> On Tue, 22 May 2018, Greg D. wrote:
>>> Oh, almost... My timezone was also wrong. It was set to -7 like I do
>>> everywhere else, instead of PST8PDT. Where did that come from? My only
>>> clue to this was that the new boot status display of time at last boot
>>> said GMT instead of PDT.
>> Because the ESP-IDF system does not have all the compiled-in timezone
>> code, we need to specify how to manage daylight saving time using the
>> long value for the TZ parameter. So not even PST8PDT is sufficient,
>> you need "PST8PDT,M3.2.0/2,M11.1.0".
>> -- Steve
>> OvmsDev mailing list
>> OvmsDev at lists.openvehicles.com <mailto:OvmsDev at lists.openvehicles.com>
>> http://lists.openvehicles.com/mailman/listinfo/ovmsdev <http://lists.openvehicles.com/mailman/listinfo/ovmsdev>
> OvmsDev mailing list
> OvmsDev at lists.openvehicles.com
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the OvmsDev