[Docs] Rework docs
Thanks @tengqm . This really is a big PR, seems like everything is refactored. I will spend time on it. Greatly appreciated.
So I took a look at the first file and there are multiple grammatical issues. Compared with the original the sentences have been broken over lines, and as previously discussed this can introduce errors. Some problems mentioned above were in the original file, but more have been introduced.
Some problems mentioned above were in the original file, but more have been introduced.
This is a ridiculous accusation. You are saying this as if you read the original docs carefully, you are saying this without appreciating the huge effort on restructuring the docs layout.
Go back home and enjoy your little pity grammar/syntax checker, or read the sembr.org for some good things to learn.
This huge pile of docs still need a lot of improvements, including accuracy and clarity. I was not pretending that I'm delivering a collection of perfect docs. There are many defects in the docs.
Is GVFS suitable to put in the client module?
cc @diqiu50 Maybe we should have the document about fuse, too.
Do we need a document java-client.md in the module client?
Do we need a document
java-client.mdin the module client?
That would be a good idea. I'd suggest we add an overview of the Java client, with references to the generated Java docs.
Is
GVFSsuitable to put in the client module? cc @diqiu50 Maybe we should have the document about fuse, too.
Currently, 'GVFS' is relocated to catalogs/fileset/gvfs, next to catalogs/fileset/hadoop.
If we have client side module(s), we may want to put it under client/.
As always, all kinds of suggestions/ideas are welcomed. For example, maybe treating gvfs as a fileset variant is not good, there is a better home for it?
Is
GVFSsuitable to put in the client module? cc @diqiu50 Maybe we should have the document about fuse, too.Currently, 'GVFS' is relocated to
catalogs/fileset/gvfs, next tocatalogs/fileset/hadoop. If we have client side module(s), we may want to put it underclient/. As always, all kinds of suggestions/ideas are welcomed. For example, maybe treatinggvfsas a fileset variant is not good, there is a better home for it?
This is GVFS client code. For me GVFS is a implementation of the client instead of server side. It wrapped the Gravitino client and hide the communication with Gravitino server. And then it implemented the Hadoop Filesystem interface. Maybe others can give another suggestion.
Is
GVFSsuitable to put in the client module? cc @diqiu50 Maybe we should have the document about fuse, too.Currently, 'GVFS' is relocated to
catalogs/fileset/gvfs, next tocatalogs/fileset/hadoop. If we have client side module(s), we may want to put it underclient/. As always, all kinds of suggestions/ideas are welcomed. For example, maybe treatinggvfsas a fileset variant is not good, there is a better home for it?
This is GVFS client code. For me GVFS is a implementation of the client instead of server side. It wrapped the Gravitino client and hide the communication with Gravitino server. And then it implemented the Hadoop Filesystem interface. Maybe others can give another suggestion.
Thanks for the clarification. My previous understanding could be wrong. If GVFS is only about client side emulation/simulation, I'm happy to find a new home for it (maybe client/gvfs?). All ears to other suggestions.
superseded by #7111
