In some cases, little to realistically assess the project in hours in advance?

When you're working freelance and doing more or less the same type of work (layout, stretch on cms, development is not very complex scripts), it is fairly easy to estimate tasks in hours, breaking them into smaller ones.

What to do if the project is complex, consists of front and backend + setup servers and we send you task represents a separate module (for example, to add a system of questionnaires).

How can we estimate in advance in hours, if you need
1)sketch of the jet components interface
2)on the backend to create the tables related to other tables
3)to develop api for data exchange, controller, model, ranting
4)to play roles, if needed
...

Just do not get to evaluate? In such cases it is necessary to break into subtasks, and to consider them as separate pieces, for example, the first point - I think the approximate number of fields that must be. Multiplied by the number of minutes spent on each get figure. But again, there is logic on the front. And need to create all sorts of functions, storage, etc.

And testing the whole bunch will deal with a lot of time. How to inform the owners? They think that once on the screen one page, then work for 1-2 hours.

And how to explain the fact that if I assessed the task in 4 hours, it does not mean I will do it so much. Most likely in such projects I have to do it every 12 hours for two days. And not going to close a day for 8 hours?

Adequately, we can only estimate the work that has already been totodiles to do. But if the task has the points that you have not previously investigated, then any such unpredictability and you're through in time. For example, if suddenly it turns out that the library, kotarou you often yuzal was incompatible with other libani project and an urgent need to find another solution, and you have voiced your client watches.

In greedy people then it's all fine. They put in 100% and throw yourself hours in advance. As do computer scientists when they, the authorities need to check average was not lower than 10 Tr
Start selling data recovery from disk polyhalite and installing drivers one by one.

And if you're humble and honest?

Here give you a job, and you know roughly how to do it, but at some point I hope to separate libraries, yet do not feel, but know that they approach this. In such cases, to be?
Or reading to include in the project cost?
April 4th 20 at 13:16
1 answer
April 4th 20 at 13:18
Well, the question definitely good. And it is connected with software engineering (at least a year at uni with that name). Donapost I see no reason to write. So here is the pickup:
1. Tehniko-the economic justification FOR
2. The evaluation model for "COCOMO"

Google, study theory. At the same time expand horizons to the level of, at least, project Manager
I suspect that the evaluation time will be measured in days. All sorts of tables, graphs, diagrams... - Missour commented on April 4th 20 at 13:21

Find more questions by tags Web Development