Evaluate Platform change For code search
What are you trying to do? Articulate your objectives using absolutely no jargon.
Evaluate if there is a better platform for code search than Elasticsearch.
How is it done today, and what are the limits of current practice?
Today we use Elasticsearch. Elasticsearch is the most used search platform in the world for Keyword searching and Log analysis.
Code search does not work like the traditional keyword search, nor does it work like log analysis. The most useful features of elasticsearch are largely disabled for code search today.
What's new in your approach, and why do you think it will be successful?
We should evaluate the Top Code search platforms to understand if they would provide distinct advantages over Elasticsearch.
- Scaling
- Repo to Index Size
- Code Search features
- Integration with Code Intelligence
- Code navigation features
Who cares? If you're successful, what difference will it make?
There are specific impacts to the amount of engineering needed and cost to operate. This should be considered along with the features enabled.
What are the risks and the payoffs?
Overlooking an opportunity means we may not enable key features. We would also miss the ability to mature code to search faster with less engineering.
How much will it cost?
The Evaluation should cost 10K. This would be accumulated in time spend and Cloud infrastructure cost for modeling. A Second evaluation may be needed and would cost 50% more to determine scaling impacts.
How long will it take?
1-3 months.
What are the midterm and final "exams" to check for success?
Phase 1
-
Determine Search platforms to evaluate -
Determine features based on UX research -
Determine the test and comparison method for non-functional evaluation.
Phase 2
-
Setup environments -
Run tests -
Determine the quality of outputs -
Lead to the total cost and feature evaluation comparison.