title | description | permalink | next_url | next_name |
---|---|---|---|---|
liq Ontology |
On how liq is defined. |
/docs/Ontology |
/docs/usage/Setup |
liq Setup |
There core resources are:
orgs
projects
work
data
services
environments
policies
- (and
meta
)
- Each project is owned by a logical
org
. - Operational policy is defined at the
org
level.
- A 'Liquid Project' is manifest as a NPM package associated to a Github project.
liq projects create
will configure each of either a new or existing NPM package and Github project.- Future versions may encode to multiple package formats.
- Future versions may integrate with other version control protocols and issues/project management platforms.
- A liq-npm
package.json
:- must include a
catalyst
(and laterliq
) entry. - must specify a GitHub hosted
repository
. - defines build and other file-type dependencies via the standard NPM mechanisms.
- defines run time through the
liq.runtime
block.
- must include a
- The current implementation does not do much in the way of checking for conformance to requirements.
- The behavior of non-conforming packages blessed as Liquid Projects is generally undefined.
- Projects define both build and runtime dependencies.
- In the liq ontology, the concept of "packages" is subsumed into the "project" concept.
- Packages are the typical method by which projects are distributed. I.e., a "package" is a "packaged project version".
- Packages are used/manipulated by the
liq projects import
,create
,audit
, and other commands
- Work represents an atomic unit of change.
- Work is managed through logical, git-based workflows and via interaction with policy touchpoints.
- The workflow ensures a basic (and configurable) level of quality control and security.
- There are three basic Environment (or Runtime) types:
dev
Environments are used to test speculative changes. They use small datasets, minimally scaled, and ephemeral.test
Environments simulate the 'real environment'.- 'User' test environments are generally moderately scaled, but otherwise fully implement a production environment right down to user involvement (a dress rehearsal).
- 'Performance' test Environments are fully automated, but simulate large scale over a short-period of time.
production
Environments are where regular users live.
- User identity and credentials passed through the Environment for use in the Runtime.
- Services, running in an Environment, create a runtime.
Additionally, there is a meta
resource which handles some basic setup tasks.