|
|
|
|
|
by timblair
4399 days ago
|
|
That's because it's the lowest common denominator. Anyone with Git on their system can use the command line and follow along. If you choose to use a GUI for your explanation, then you're likely alienating at least a portion of your readerbase who can't use / don't want to install that interface. I agree that a GUI can make some tasks a lot simpler but, personally, I've always found that teaching people to use any DVCS from the ground up (i.e. from the command line) ends up with a greater level of understanding. Once you've got that, you can move on to using a GUI to improve your workflow, but with the understanding of exactly what is being abstracted away. |
|
In any case, I'm quite sceptical of the claim that you can teach Git more effectively through the command line alone. Here's why: the two most important concepts to understand before you can properly grok DVCS are (a) that your source history is a key integral part of your workflow and not just a sideshow, and (b) how revisions, changesets and merges relate to each other through the DAG. GUI tools do this very effectively by putting your source history, complete with a graphical view of your revisions, right at the front of your workflow. By contrast, the command line requires you to type git log --graph, effectively relegating it to a sideshow like in the bad old days of Subversion.
I can't help getting the impression that learning DVCS through the command line risks you ending up with an understanding of the subject that is fundamentally flawed. You see this in virtually every Git versus Mercurial debate that focuses on which one is better at branching and merging. Most of the arguments that I've seen leave me with the impression that the people who are making them haven't a clue what they're talking about.