I'd recommend the 5x5 method which discussed on the in detail on the critical-hits.com blog in the context of both adventures and campaigns
The basic idea of this technique, is that you take 5 ideas:
- Overthrow the evil empire
- Prevent the sorcerer from ascending to demon-hood
- and 3 more...
and so on, and then for each idea create 5 milestones that will happen along the way.
e.g. for overthrow the evil empire
- meet up with the rebel alliance
- learn the weakness of the death star
- and 3 more...
and for prevent the sorcerer from ascending to demon-hood
- steal the orb of power from the temple of evil
- kidnapped by the sorcerer's lizard man minions
- and 3 more
Put that on a grid:
Overthrow the evil empire Prevent the sorcerer from Thing 3 Thing 4 Thing 5
ascending to demon-hood
learn the weakness of the steal the orb of power .... .... ....
death star from the temple of evil
meet up with the rebel kidnapped by the sorcerer's .... .... ....
alliance lizard man minions
.... .... .... .... ....
.... .... .... .... ....
.... .... .... .... ....
And then using that as scaffolding build around them.
One of the nice aspects of this method is that you can give your players control over the paths they take. They can chose to ignore one of the lanes that you have sketched out without derailing you completely. And you can spring natural consequences of their actions on them, that make sense in the context of what has gone before.
critical-hits.com also has a summary of other places on the internet that talk about this idea
Most campaigns don't reach their end
That's just the way it is. Doubly so for your first ever campaign. You might lose interest. So might your players. You might realize you don't know what to do with them anymore. Life might intervene. Things happen.
And that's ok. Fun would still have been had. Memories would still be formed. The world you create might become a place for a next campaign - or inspire you to make a new one. As banal as it sounds, you'll learn from your mistakes (and discover exciting ways to make different ones). Don't sweat it.
Here, then, is some general advice on running a lengthy campaign.
Break it up into acts
Knowing as you do it's likely to prematurely end, plan for that. Have several major acts that would provide some satisfying conclusion. 4e's tiers work well here, though it does depend on the pace of your game - how quickly the PCs will level.
Perhaps in the first act, the war with orcs is a straightforward affair - they invade human lands and defeating their force is the end goal. In the second act, paragon-tier PCs venture into orc lands instead, to discover how orcs have changed. They also witness humankind begin its downfall, and make a decision in the end on which side to favor. In the third act, epic-tier PCs engage with the mystical source of knowledge and civilization (God? Artefact?) and rewrite the nature of man and orc. At that point, mortal kings are beneath their concern, as they deal with cosmic forces themselves.
Don't save the cool things for later
Do you have an awesome idea? Use it now! Don't store it until the slightly more appropriate time two years from now. By then you'll either forget it, it'll become irrelevant, or you won't get there at all. Make each session fun, not a prelude for fun down the road.
Plot it out
You have an idea for what you want the campaign to be like. Use some of the tools developed by others to help you get there. 5x5 method or Dungeon World's fronts, for instance. It won't do you any good holding it all in your head - get it out on (virtual) paper.
Change it up
Those methods tend to say this, but it's worth repeating: if the game goes elsewhere - that's fine, too. Never be afraid to change your plans. Don't stick to the one true path.
Play the same game
This is what kills most campaigns - people play different games at the same table. Use The Same Page Tool to set common expectations. And always, always, always talk to your players.
Best Answer
If you're interested in the geographic scale of a game world, then you need to consider the following points.
On land at least, settlements tend to be at most around an easy day's travel from each-other, to allow for people to travel to the next town without having to stress themselves or their beasts of burden. At sea, the equation changes as ships can carry their own supply.
Considering a Pathfinder game, we would expect that, barring the relatively rare exception of magic or exotic creatures, most transport will be muscle-powered, i.e. walking, riding animals (e.g. horses) or riding in animal-drawn conveyances on land, and on water, boats and ships will typically be wind or muscle powered. The following links provide some reference and discussion as to typical and exceptional travel times in this type of environment:
http://www.cartographersguild.com/reference-material/19730-how-far-horse-travels-one-day.html
http://www.enworld.org/forum/showthread.php?318719-Travel-times-in-fantasy-pre-industrial-society-by-foot-horse-boat-etc
http://paizo.com/pathfinderRPG/prd/additionalRules.html#table-7-6-movement-and-distance
To summarise the relevant data from the above in case of link rot:
From this, we can conclude that, depending upon terrain and the quality of the path between settlements, settlements could be anything from 10 to 40 miles apart, though the highest likelihood, assuming that roads have been built, is 20 to 30 miles apart.
You should consider that there won't be cities every 25 miles on average, however. Most of these settlements are likely to be some sort of village which serves as a collection point for agricultural produce and as a way-station for travelers who can't afford to use more exotic means of transport such as magic or flying mounts. Where the logical routes of travel dictate that more commerce will pass through a single area, the settlements become larger. Other factors such as water supply will dictate how large such settlements can become.
In a world with flying mounts, we can expect that some of the settlements that would otherwise be villages could be larger to accommodate riders of the most common type of flying mount - typically within a day's easy flight of each other. However, unless a region was difficult to fly through (extreme mountains, chronically bad weather), the spacing of these would not really depend on terrain. Also, there may well be some smaller settlements in areas difficult to access by land that exist as way stations for flying travelers.
Extrapolating from Pathfinder's Movement and Distance table, some example flying mounts:
Spells/items: - Overland Flight/ broom/cauldron/carpet of Flying / Flying Ointment (variable)
You would need to decide which of the above (or other) was in common use - probably overland Flight spells/items, or hippogriffs, as they are the most easily obtained, though allowance should be made for slower, rarer creatures such as griffons or even wyverns if such creatures are used much. Way-stations for flying creatures/magic users will typically be placed within a day's flight for the slowest in reasonably "common" use.
Finally, you need to consider the size of sovereign nations. These will be dependent upon many factors, and barring a war in progress, a nation's borders are typically tied to some sort of natural boundary such as mountains, rivers, woods, coastlines, etc. The size of a nation depends on its ability to patrol and defend its land, and in turn this will depend on the strength and belligerency of its neighbors - the stronger and more belligerent the neighbors, the more a nation will have to concentrate its defenses to counter them, and will consequently be less able to defend its interior and other borders.
As to the size of your entire world, that is entirely a matter of personal choice. You needn't even define your whole world to begin with, just your preferred area and name a few mysterious far-away places.