UX/UI issues & enhancements
This is a running compilation of notes taken over the last month I've been working with pimp. There will likely be redundancies in the text.
As a background note, I have worked in digital production for 20+ years with a special in large-scale content management systems and the tooling (UX/UI) of such.
Pimp is a fantastic tool and I would like to help out with any improvements. If the team ever intends on making changes to miner.farm, let me know and I can contribute wireframes and the like.
- After session timeout clicking from a view to pool view gives a bad gateway error, requires page refresh. Session getting lost, and AJAX not setting a true logged out state
- Default redirect to profile page after logging in has little to no use. Redirect to farm view by default or have a profile setting that specifies what your preference is. Based on my main body of suggestion how to integrate views, just redirect to farm view instead.
- The errors per hour are convoluted by the difference (or lack thereof) between warnings and errors
- Warnings occur when a miner is off, but a miner cannot necessarily run if another miner is already running. There should be a way that a non-running miner can be set to NOT incur a warning and make the errors screen convoluted. I also question what the difference between a warning and error is treated, as the errors bar is where both are reported. I want my email alerts to warn me when I actually need to do something and so warnings that a miner is off-line is really a non-issue. When my miner has a GPU going sick or dead, that is a warning I want to get and action, and should actually be an error, not a warning/
- Based on above I suggest that alarms can be set on different entities, a farm (agent), a pool and a miner
- An alarm logic would be useful where the results of the data being monitored can cause action. For instance, if a miner has GPUs that are sick or dead or HWs go above x%, miner is restarted and log a warning. This could be expanded to cumulative logic as well, if X warnings are logged within a period of time, the box is restarted
- General nesting could be enforced more. When a piece of data belongs to a parent, it gets tabbed. The IA generally shows data following, farm view shows agents, with miners following. These could be nested more to show miners belonging to the agent, which belongs to a farm. Controls could be incorporated into this IA, hence the comments below are mostly about merging screens together. Restart controls especially should be incorporated directly in-view.
- Pool is a really sub of miner except that you identify any and all pools that may be configured in the miner.
- Without being able to "switch pool", I believe that the pool view really has no use right now
- Hence the pool view is really a report of what is the currently active pool which is tangential to the miner itself, you can easily merge both views and a miner has an expand/collapse function which opens to reveal the pools, and a feature addition can switch active pool
- Alert view only shows miners that belong to a profile. If you could apply alert logic to agent, miner, pool you can make this view nested based on the hierarchal relationship between agent, miner, pool which it's individual reporting logic
- Pool view even more so, it shows crawled pools irrespective of it's relationship to a miner or agent, in which case it's a clusterf### of data which doesnt really tie to anything. Considering you cannot action what this does or mean, it seems redundant. This view could come back to life if you could by-control switch pools. In which case you would again, present the data by nesting it based on its hierarchal relationship between agent, miner, pool.
One this I would def add on the O/S side is that something should be built to allow any intel onboard graphics to be "ignored:" by pimp rather than insist it be disabled. this makes little sense when you dont want to drive the x-term using one of the GPUs for mining.
Thanks Redfred5. Well thought out and described. We have started building PiMP & Miner.farm 3.0, and these kinds of ideas will be driving development. When we have a beta, we will most certainly include you if you wish..
Catch22 last edited by
So many good point in this post. Glad to here the team is already working towards some of these. One thing I'd like to add in is on Miner.farm when I go to the alerts view I get a different GPU count on one of my rigs than on the farm view. I only have 8 GPUs on the rig and Miner.farm shows that when you look at the rig details but on the alerts view it shows that I have 9 GPUs. May something to do with this?
Agent Host GPUs
GPU 01: AMD/ATI Radeon RX 470/480/570/580
GPU 02: AMD/ATI Radeon RX 470/480/570/580
GPU 03: AMD/ATI Radeon RX 470/480/570/580
GPU 05: AMD/ATI Radeon RX 470/480/570/580
GPU 06: AMD/ATI Radeon RX 470/480/570/580
GPU 07: AMD/ATI Radeon RX 470/480/570/580
GPU 08: AMD/ATI Radeon RX 470/480/570/580
GPU 0a: AMD/ATI Radeon RX 470/480/570/580
Skips 04 and ends at 0a so maybe that is where the 9 comes from? Just something I noticed.
@catch22 However, they are still 8 :) Slightly broken addressing, you can see how they look with
Catch22 last edited by
@blackice True, I know I didn't gain a GPU (I wish I did!) just another weird little bug because that means the Farm view and the Alert view are generating the total count in two different ways since they get two different values.
@catch22, yes please. i certainly would like to be involved. i started wireframing in 2000, so ive done literally thousands of them.
Guys! For an awesome chat experience and great mobile and voice support, come talk to the PiMPs on Discord! Also it is connected to the IRC so you can be in touch with everyone: https://discord.gg/grCbFsA
We promise to keep your email safe and never spam you.
Copyright (c) 2017 PiMP LLC. All rights Reserved.