How do you estimate the time required for a programmer to solve a problem?

Question

JIRA

A purely domestic issue for small businesses. For example, I've never written software for Windows in my life, and it's hard for me to appreciate the complexity of solving the tasks. Probably worse only for those who have never programmed at all.
For humanistic reasons for the developer and myself, I want to simplify the "estimation" process as much as possible so that it is simple enough for a layman and fast.
With a correct assessment of the work, the advantages are obvious:
- the employee does not imitate violent activity
- the employee is not overloaded
- the chances of deadlines are reduced
What to read in relation to small businesses, when there are simply no intermediaries, for example, in the form of management and senior developers.
What formula do you use for this at home? How do you reproduce knowledge about aspects of activity?

Answer 1

In general, there are formulas for calculating terms, but how effective they are is a big question. If you are interested, google on the topic "Broadband Delphic method", "PERT"
We usually practice 2 methods: the developer indicates the number of hours, how much he will complete the revision - for example, from 4 to 8, the manager writes to the manager 6-10.
Another method - the manager said - we will do it in 3 days, the developer can do it or not - his difficulties. We usually have time, but after that the code becomes dirty, then it is not very good to maintain, programmers are extremely dissatisfied with the work of the management and managers.
If the implementation requires the study of new technologies, libraries - this is no longer an indication of the timing, but fortune telling. From a month to three, for example.
The concept of task readiness also depends on the one who indicated the deadline. If the deadline was specified by the programmer - this is development + internal testing, if the manager - the deadline for launching the revision.

Answer 2

My applied method for estimating time costs is as follows.
1) Make the maximum decomposition of the task (break into subtasks) to such an extent that each subtask is either already performed by you, or you have a clear idea of ​​​​how to solve the subtask.
2) Make a list of risk factors in the form of multipliers. For example: a new platform - 1.2 from the base, a complex client - 1.5 from the base, etc.
3) Multiply the subtasks by the risks that apply to them, then add up the results.
This approach gives me a variance of 10% of the actual cost.

Answer 3

The estimate on time should be given by the one who will do this task.
Moreover, at the time of the assessment, he should more or less imagine how he will do this task.
Otherwise, this is not an assessment, but guesswork.

Answer 4

it is better to always budget the time budget for the task with a 20% or even 30% margin on top.
all the smart books on software development say that in most cases the project will never be completed on time if you set a hard deadline. from experience and practice. deadlines are always postponed and appropriate decisions and agreements are signed. Of course, it also depends on the number of people involved in solving the problem. the degree of risk of non-fulfillment of the temporary budget is directly proportional to the number of project participants.
if, as stated in the question, only one or two performers appear (for small businesses), then the work and time budget can already be estimated more accurately. but again, lay in excess stock, force majeure, etc.
a lot depends on the accuracy, completeness, clarity of the requirements for the project being developed or the task being formulated, descriptions of design solutions and their consistency, understandability for performers.
during the execution of a task / project, check the degree of readiness at some control points, which will more protect against overspending of the time budget