Talk:Improved Fuel Expenses Calculation

From RadiusWiki

Tank capacity

Nikolai:

I have some concerns about fuel consumption calculations - after fuel tank volume will be introduced, will it still be allowed to fill more fuel than the tank volume?


Radius:

No, I don't think so. Do you? I think it's impossible to tank more than the capacity really is. If you use a gas can too, the filling should be split into 2 records, I think.


Nikolai:

Yes, I do. I have a long statistics for my car (imported into RadiusFuel from another program) and I used to combine data for consequent non-full tanks up to the next full tank.

For example, my tank is only 68 l, but I have a lot of fills like 71.54 l, etc., which are actually 2 fills - 10 l and 61.54. And I don't have info for any of these fills - only for the sum.

That means, that I will not be able to import all my previous statistics into RadiusCar :(

Frankly speaking I didn't analyse this section yet :( My first idea would be to allow to fill more, but to ask user for confirmation before saving the fill info.


Radius:

Well, everything what helps you and what is useful will be implemented if possible. Ok, RadiusCar needn't limit filling amount if it's not desired. But please think about your proposals at Improved Fuel Expenses Calculation. I think we may encounter some problems in our costing models while allowing "overfills".

Thinking about your statements above I have to change something... I have already implemented a part of RadiusCar and it now allows to automatically import RadiusFuel's data but it asks for your car's tank capacity. It's ok up to now. The problem is RadiusCar does not allow tank size less than the biggest fill ever was. I should change this behavior as well, right?


Johny:

Radius,

I believe you should skip (or let optional or have some tolerance) control of full tank. It has happend to me also (official tank capacity 55l - Skoda Octavia) and I have pumped 57 liters. I believe there is difference - tank capacity is only real tank, while you can have petrol also in tubes.

Johny


Nikolai:

Hi!

First, I agree with Johny - RadiusCar should be flexible and allow user to fill in more than the tank capacity is. Nevertheless, RadiusCar should analyze this situation and ask a confirmation from user - a dialogue window like "You entered more fuel than the tank capacity of your car. Accept/Correct" (we need a native speaker to reformulate this question in a better way).

Second, I would like to propose minor corrections into "easy mode" of fuel consumption calculation. In this mode a check-box should exist - "empty/non-full/full". If user chooses "non-full", no consumption calculation should be done, only fuel price per liter etc. This fill value should be added to the next empty/full tank. This approach will allow really exact calculations of fuel consumption.


Radius:

To Johny: I am very surprised you reached the gas station in this case. It seems your car runs on pure air too ;-)

To Nikolai: Well, RC will allow overfills on explicit user's confirmation. This seems to be a correct way how to retrieve data. I will try to create an example of overfills to make our costing models to become invalid. It seems possible to me. Nikolai, please add the question on overfill to the specification in the main article. Thanks.


Nikolai:

Hi!

Well, Johny mentioned actually another problem with overfills.

Basically we have 2 cases:

  1. You can fill in more fuel than your tank volume is.
    For this situation you need to include overfill question to user interface and (probably) correct a bit costing model.
  2. You can combine several non-full fills into virtual one fill to make exact fuel consumption calculation.
    This was actually the only possibility in RadiusFuel and should be supported in RadiusCar as well both for backwards compatibility and for paranoia users (like me :)).

To support this, a special costing model for easy mode should be introduced, I believe. Maybe you already got the idea, nevertheless, let me try to explain it below:

Easy Mode costing model

Nikolai:

I'll try to make the explanation based on an example.

Imagine, you always used to tank up to full-tank. Than a simple "RadiusFuel" costing model works very well and reliable, giving you real fuel consumption every fill. You even do not need to know the tank volume actually. I believe, Easy Mode should simply continue to use this "old good" approach with only one useful extension - overfills or virtual fills.

If you, by way of exception, will not be able to make a full-tank fill, filling in, by example, only 10 liters, you will broke this simple costing model and the fuel consumption for this fill and for the next one fill will be wrongly calculated.

To make it good, I would propose the following approach:

  1. RadiusCar provides in Easy Mode additional checkbox full tank/non-full tank, which is by default set to "full tank".
  2. If user fills in less than full tank, he or she should check "non-full tank", RadiusCar should calculate only fuel unit cost and provide following info in the main table view only
    • Record number
    • Car identification
    • Date of expense
    • Amount (fuel volume)
    • Cost of expense (cost of fill in this case)
    • Unit cost
    • Odometer
    • Consumption (no value!!!)
    • Endurance (no value!!!)
    • Station (tank station)
    • Type of expense (Fuel)
    • Comment field
      Please note that Consumption and Endurance values should be empty in the table!
  3. If user makes a full-tank fill after one or several non-full tanks, RaduisCar should calculate internally a kind of virtual fill by taking a sum of all previous consequent non-full fills as a basement for Consumption and Endurance calculations as well as for Fuel cost per 100km, Distance per day and similar. Such a virtual fill can exceed the actual tank volume!
    In the main table view we will see the following:
    • Record number
    • Car identification
    • Date of expense
    • Amount (fuel volume)
    • Cost of expense (cost of fill in this case)
    • Unit cost
    • Odometer
    • Consumption (combined value for this fill and previous one or several non-fills!!!)
    • Endurance (combined value for this fill and previous one or several non-fills!!!)
    • Station (tank station)
    • Type of expense (Fuel)
    • Comment field
      Please note Consumption and Endurance values again!

In this model graphics for Unit (Fuel) cost, Cost of fill, Fuel cost per day should take all the real fills into account, but Consumption, Endurance, Fuel cost per 100km and Distance per day graphics should use virtual fills instead of non-full tanks.


Using the RadiusFuel application, I had to calculate these virtual fills by myself, deleting the non-full tank fills and creating virtual overfills. Thus I have correct statistics for consumption etc., but do not have the information for non-full tanks anymore... and have potential incompatibilities with RadiusCar, which should be avoided :)...


Radius: I agree with you, Nikolai. So I copied your article above to the main page of fuel expenses with the only exception: I do not think Distance per day should be calculated from virtual fills. Please explain me why... Distances and dates are correct values anytime.

What do you think about showing bad calculations based on non-full fills, but displaying them with exclamation icons (or something similar) instead of empty cells in the table? I think some approximation could be used.


Nikolai:

Distance per day:
The explanation is very simple and obvious - it is just my mistake. I am ashamed of this. Well, to be serious, I was to concentrated to present you the general idea and simply overseen this fact.

Bad calculations:
May I ask to make this feature "switchable" through preferences dialogue? :) I like your idea actually.

Full tank/non-full tank checkbox in the example above:
Above I presented more an example than "specification". In general case, the same approach applies also to "empty-tank-fills" in Easy Mode. So, depending on the general preferences you mentioned in the article, checkbox should present either

  • full tank
  • non-full tank

or

  • empty tank
  • non-empty tank

Johny:

Hi,

I would like to intervene into your bad calculations discussion. If you want to show bad number with exclamation icon, I would propose to use for consumption and endurance some kind of average or last correctly calculated number. It is in my view much closer to reality (less bad) than number calculated RadiusFuel way (I would call it very very bad, since I am sure it is not good). :-)))

Johny


Radius:

Hi guys,

I changed the main article to reflect your proposals. RadiusCar should provide as accurate calculations as possible. The user will see inexact calculations displayed in a special manner.

Expert Mode Costing Model

It seems to me, the expert mode is not as important as we thought. Is it really possible the driver will know exact percentage of fuel remaining in the tank just before filling? It sounds like science fiction to me :-) What do you think about removing the expert mode?


Nikolai:

I do not think, that you should cancel expert mode. I know a lot of people, who never fills up to full tank. For them expert mode will be much more interesting (despite the fact, that it is not precise).

Actually, I do not like easy/expert names - easy mode is much more difficult to achieve (you cannot make full fills every time), and expert mode is more understandable for ordinary users :) At the same time, easy mode is precise, while expert is always approximate...

You should probably think about different naming :)