The perennial conventional wisdom is that a new writer ought to learn to write short stories before tackling a novel. The theory goes that working on shorter lengths will allow you to master various aspects of prose and storytelling craft while giving you the satisfaction of actually completing a story in a reasonable amount of time. We all need those gold stars, right? the more so when we're struggling to learn something new.
Another argument for "short first, then novel" is that you can establish your professional chops by selling to magazines and anthologies, and thereby achieve the name recognition that will help get your novel read or represented. When your novel does come out, you'll have a reader base.
This strategy has certainly worked for many writers in the past, and undoubtedly will work for many more in the future. That's because most -- but not all -- writers are creatively-wired to work at different lengths. It was and still is easier for most -- but not all -- writers to sell a short story than a novel. Let me elaborate:
The theory of learning to construct stories by writing shorter lengths (if we extend the category to include novelette, up to 12,500 words or about 50 manuscript pages) presupposes that we can thereby focus on only a few elements at a time. There will be one major plot line and relatively few characters, and the setting will not require great elaboration. This is an overgeneralization, of course. Short stories don't often feature an array of subplots and cast of thousands (unless they're nameless hordes).
What's wrong with this argument is that short stories by their very nature are compact, as opposed to the expansiveness of a novel. Every detail, every element must do double or triple duty (for example, a line of dialog might reveal character, advance plot, and evoke the world or society or family relationships at the same time). While it requires concentration to juggle multiple subplots, it is not proportionately more difficult than depicting a single line of action. Characters still need to be well done; settings still need to be specific rather than generic; dialog still needs to have certain characteristics, etc. The difference is that you have to do all that in, say, 5,000 words instead of 100,000 words.
Another problem with the conventional wisdom is that the print market for short stories has shrunk dramatically. Dramatically is too tame a word. Very few anthologies are being published nowadays compared to a decade or two ago, and of those, even fewer are open to unsolicited submissions. Instead of 20 or 30 paying magazine markets, we have a handful. So the new writer who thinks it's easier to sell a short story may be sadly mistaken when he finds himself in competition with multiple award-winning authors for the same few slots.
Perhaps the most powerful objection to the short first, then novels dictum is that one size does not fit all. Some writers are natural novelists and struggle painfully to learn to write at the same time as trying to master a length that doesn't make intuitive sense to them. Once they're seasoned professionals, most of them can write short if they have to, but without the ease of their preferred length. Other writers naturally think short; it's how their minds work, and I've always thought it a shame to waste that brilliance forcing it longer and longer until all the joy has been stretched out of the story.
The explosion in epublishing promises to bring some interesting changes to the field. Now we have magazines that are entirely online or downloadable, without any print versions at all, and others that offer both. Established authors are bring out not only their novel backlists but their out of print short fiction in electronic form. Original works at both lengths are widespread. Short fiction is especially attractive to people who prefer reading their stories on handheld devices; short stories are perfect for airports and waiting rooms, offering the satisfaction of being able to finish the story in a single sitting. The low prices of short stories add to their appeal. Being able to purchase and download stories individually allows the buyer to tailor-make her own anthologies. It's entirely possible that these developments in publishing technology will lead to a renaissance in short fiction as an art form. I hope so, but I wouldn't count on it to pay the rent. At least, not yet.
Harry Turtledove once said that novels teach you what to put in a story and short stories teach you what to take out. What's left, at its best, is a jewel whose every facet is precise and clean-edged, a thing of glory to read and wonder to write. A great short story packs a different kind of punch than does a great novel, and one does not substitute for another.
So why write short? For me, there's only one good reason (aside from being powerless over what the Idea Fairy leaves me). That's because I love the form. I love writing it, how it makes all the parts of my creative brain go into hyperdrive coordination, that zing! when everything I've set up crystallizes, the high wire act of having to get it exactly right. I love reading it because I'm in awe of a story that excels in doing so much in so few words.