design review cleenup algorithem

Intro:

Areas change through the years, the city develops and living consition change.
Our users want to know what is the state of the area by an apartment they are checking out.
It would be beneficial for the users to have accurate information.

---

we have diffrent aspects of this issue to handle:
* How to know if a review is still relevent?
* How to remove an irelevent review?
* How to implement the removal process?

Relevency

we have two options in this part:

1. All reviews are relevent for ever
2. Reviews become irrelevent due to time passing, natural changes and unnatural changes (man made changes)
   

If we choose the former, this feature is done in the empty sence of the word, nothing to do, we can go home. obviously, we will choose the latter due to us wanting to give good and reliable service and information.
Now what is left is to decide how to "know" a review is irrelevent, how fresh does the review need to be?
a year seems like a good choice, due to our user demographic being of roomates finders, but we still have to deal with changes we cannot predict. We will change a review to irrelevant manually.

Removal of irrelevent reviews

We can go in a number of ways:

1. Remove the review from the database completly, both from reviews and from the poster's data (list of reviews postes by a user).
   
2. Sign the review as old, display it in the user consule as inactive, this way we can give the user the option to revive the review if he sees it as fit.

We will implement the feature with the second option.

Implementation of the removal process

When to remove?

1. We can do this on the fly, when a user asks for a review we will first check the area if any review became Irrelevant since last checked and then calculate the desired review.
   
2. We can also start the process at scheduled intervals, once a year, once a month or once a day all together.

The first option will delay the user, but the second might acutaly freeze the feature temporarly each time we start the removal process due to the number of reviews we hope to obtain in the future.
Another very importent point is if to remove the last review in a redius of 1 km, which is the radius for review calculation, leving an area reviewless.
Seems we will have to have a minimum number of review in an area that we will not remove. We will need to rank the reviews to get rid of the most Irrelevant reviews.

decision

We will show the Irrelevant reviews in the user consule as inactive as stated above.
A review will become Irrelevant after 2 years and a year later will be deleted completely.
For now we won't have a minimum due to the difficulty of the problem.
for now the cleen up we accure on the fly as people ask for the calculated review