Opinions about tilemaps vs. sprites for platform environments

I use Construct 3 in an undergraduate communication course focused on game design and criticism. Once students understand the basics of the application, their first game design assignment asks them to create their own platform game.

In the past, following the recommendations of the original Construct 3 tutorial, I have encouraged students to create islands, ladders, and other elements of the game environment by loading a tile sheet with environmental elements into a single sprite. Each environmental object is a different frame of the sprite, the speed of the animation is set to zero, and students design the environment by dragging sprite instances around the layout and selecting frames accordingly.

However, I noticed that the most recent platformer tutorial on the Construct site suggests using the tilemap object instead of using sprite instances. If this is the preferred workflow, I will encourage my students to use this approach, but I'm wondering what others think about this.

What are the advantages or disadvantages of using tilemaps to create the environment in a platform game? Is it just a case of "six of one, half dozen the other?" Or are there clear advantages to one approach?

Its whatever is best for the job.

If you want to use up a lot of area with a texture, then Tiled BG is better than Tilemap.

If you have lots of platforms, or roads, or buildings, then Tilemap may be better than Sprite.

Thanks, Newt! You mentioned that Tiled BG is better for areas with texture and Tilemap is better for lots of platforms, roads and buildings. Is the only noticeable difference in the creation phase when the developer is designing the layout? Or are there also differences related to code and ease of collision detection?

Good question, I would like to hear from scirra's team which approach is recommended.

There are some considerations like:

Which approach does have the best performance and why..?

Should I use tilemaps, Should I use a single sprite for each object, Should I use a sprite with multiples frames for each object?

Which approach is more likely to be the industry standard?

Tilemaps are better for performance than Sprites, it explains it in the manual under Tilemaps.

If you can use a tilemap, you probably should. However tilemaps are limited to same-size images in a grid. You might still want to use the technique you described (sprites with different initial frames) if you want more flexibility, such as positioning different sized objects freeform, not aligned to a grid.

Using tilemaps is faster than using a grid of sprites to produce the same effect. But other than that, it's probably too soon to ask questions about performance - and if you want a good answer, you can measure the performance yourself.

Thank you, everyone. This is exactly what I needed to know!

I should mention mesh's. They offer a few features that Tilemap does not.