From: "Karsten Blees" <karsten.bl...@gmail.com>
Am 15.10.2013 00:29, schrieb Felipe Contreras:
tl;dr: everyone except Junio C Hamano and Drew Northup agrees; we
away from the name "the index".
It has been discussed many times in the past that 'index' is not an
appropriate description for what the high-level user does with it,
it has been agreed that 'staging area' is the best term.
I haven't followed the previous discussion, but if a final conclusion
towards 'staging area' has already been reached, it should probably be
Do you mean that how that conclusion was reached should be summarised,
or that you don't think it's an appropriate summary of the broader
The term 'staging area' is more intuitive for newcomers which are
familiar with English than with Git, and it seems to be a
straightforward mental notion for people with different mother
In fact it is so intuitive that it's used already in a lot online
documentation, and the people that do teach Git professionally use
term, because it's easier for many kinds of audiences to grasp.
Such online documentation often portraits the 'staging area' as some
supposedly 'unique' git feature, which I find _very_ confusing. In
fact, every major SCM has a staging area. E.g. you first need to
"svn/hg/bzr/p4 add/remove/rename/move" a file, which is somehow
recorded in the working copy. These recorded changes will then be
picked up by a subsequent "svn/hg/bzr/p4 commit/submit".
Additionally, all those systems support selectively committing
individual files (by specifying the files on the commit command line
or selecting them in a GUI).
So git's 'unique staging area' boils down to this:
1.) Recording individual files to commit in advance (instead of
specifying them at commit time). Which isn't that hard to grasp.
For many, that separation of preparation(s), from the final action, is
brand new and difficult to appreciate - it's special to computer systems
(where copying is 100% reliable, essentially instantaneous, and in Git's
case, 100% verifiable via crypto checksums).
2.) Recording individual diff hunks or even lines to commit. Which is
an advanced feature that requires even more advanced commands to be
useful (i.e. stash save --keep-index; make; test; commit; stash pop).
It is also entirely irrelevant to binary files (i.e. for non-technical
users that use git to track documents and stuff).
index: an 'index' is a guide of pointers to something else; a book
index has a list of entries so the reader can locate information
easily without having to go through the whole book. Git porcelain is
not using the staging area to find out entries quicker; it's not an
The 'staging area' is a sorted list of most recently checked out
files, and its primary purpose is to quickly detect changes in the
working copy (i.e. its an index).
There is a big (human) problem here. We (humans) are able to believe
contradictory things ("He ain't heavy, he's my brother" to quote a
song). The Index (file) isn't a staging area, but we are happy to flip
flop between the two ideas depending on context - others can feel
In one sense the "Index" is an implementation detail of the concept of a
packing area where a shipment (commit) is prepared, which is most
commonly called the staging are in populist discussions (which I believe
is the summary I mentioned above)
Of the 28 builtin main porcelain commands, 18 read the index
(read_index), and 11 of those also check the state of the working copy
(ie_match_stat). Incidentally, the latter include _all_ commands that
update the so-called 'staging area'.
Subversion until recently kept the checked out files scattered in
**/.svn/text-base directories, and many operations were terribly slow
due to this. Subversion 1.7 introduced a new working copy format,
based on a database in the root .svn directory (i.e. an index),
leading to tremendous performance improvements.
The git index is a major feature that facilitates the incredible
performance we're so much addicted to...why be shy about it and call
it something else?
stage: a 'stage' is a special area designated for convenience in
for some activity to take place; an orator would prepare a stage in
order for her speak to be successful, otherwise many people might not
be able to hear, or see her. Git porcelain is using the staging area
precisely as a special area to be separated from the working
I'm not a native speaker, but in my limited understanding, 'staging'
in computer jargon is the process of preparing data for a production
system (i.e. until the 'stage' or 'state' of the data is ready for
production). It has nothing to do with the 'stage' in a theater. I've
never heard the term 'staging' beeing used for source code or software
(that would be 'integration'). I've also never heard 'staging' for
moving data back from a production system to some work- or development
Even 'native' speakers don't have a single consistent term for the
concept. Terms are stolen from many varied industries and activities
that have to prepare and package items (Ships, Trains, Theaters)
(see http://en.wikipedia.org/wiki/Shipping_list, for a shortish list,
which doesn't mention an Index)
In any sense, 'staging' is a unidirectional process (even in a
theater). If I think of the index as a staging area, it covers just a
single use case: preparing new commits. With this world view, it is
completely counter-intuitive that e.g. changing branches overwrites my
IMO, it is ok to use 'like a staging area' when we talk about using
the index to prepare new commits. However, its not ok to use 'staging
area' as a general synonym for the index.
As alluded to above, the developer viewpoint includes a perception of
how we implement the preparation of commits, and how 'we' have come to
fix on the term 'index'.
In one sense even that is not the right term - If compared to a book /
pamphlet / monograph (being placed in a Library / repository) it's more
of a contents list (by chapter and verse / directory and file), with
various bits of front matter such as author, publisher, previous
editions, introductory preface, dates, contents list, and finally
content. A book's 'index' is a supplementary mini grep of useful terms
that the reader may wish to find.
All in all it's difficult to undo this Gordian knot of confusions.
The key is probably to separate the devlopers concerns over
implementation details from the user's big picture view, in an arena
that is short of well (commonly) understood terms.
Just my 2 cents
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html