The main factor that increases the size of your help files is the use of big graphics with too many colors. Compared to images text hardly takes up any space at all.
In many cases you may find it most practical to use standard uncompressed BMP bitmap files for your graphics in your projects. This may seem odd at first: BMP files are generally larger than all other graphic files, as they are without any compression at all. However, they can be converted most easily and since they are uncompressed they suffer no quality degradation. Help+Manual converts BMP files to the appropriate format and compress them automatically when you publish your output. Like the XML format used for the source of your topics, BMP is a flexible choice for exporting to multiple formats.
If you edit your images in Help+Manual's Impict graphics editor you will often add shapes, text callouts etc. If you save these images in Impict's own .IPP format you will be able to re-edit all the objects you add later. You can insert Impict .IPP images directly in your topics. They have lossless compression and can be converted to all other formats required, just like BMP.
Here are some basic rules of thumb for minimizing the size of your graphics and output files if you need to keep output sizes to an absolute minimum. However, file sizes are really not as important nowadays as they used to be. You may find it more practical to relax these rules and use full-color BMP and IPP images across the board.
This will generally reduce your output file size more than anything else. You will almost never need more than 256 colors for screenshots unless your program contains graphical components with complex color gradients.
Help+Manual allows you to insert compressed JPG images directly into your projects, along with many other graphics formats. However, JPG images are always TrueColor, so it's not possible to reduce the number of colors to 256 or less. Also, the compression used in JPEG images often makes screenshots look bad because it creates ugly "artefacts" at the sharp transitions that are typical in elements of items on computer screens.
It's better to use uncompressed BMP images with 256 colors or less and allow Help+Manual's Image Conversion function to handle the compression at publish time. If you want to do the compression yourself use GIF or PNG for 256-color images, the quality will be much better than with JPG.
The Image Conversion settings are in the HTML Export Options, which can be accessed in both the HTML Help and the WebHelp sections of Configuration > Publishing Options in the Project Explorer.
The PNG format is an excellent choice for images with up to 256 colors. It is even more compact than GIF and has excellent quality. However, PNG files with more than 256 colors are actually a different format. These files are significantly larger than the equivalent JPG files. Sometimes they are only a little smaller than BMP files.
This also means that you should make sure that your image conversion settings will not convert images with more than 256 colors to PNG. If your project contains images with more than 256 colors it is better to choose JPG as the conversion format.
Modern graphics programs can do an excellent job of reducing photos to 256 colors with little or no reduction in visible quality, which also saves space. However, doing this can cause a problem in Help+Manual if you save the images as BMP.
If your image conversion settings are set to convert images with 256 colors or less to GIF any photos with this number of colors will also be converted to GIF. This can make them look bad, with nasty posterizing effects, and it can also increase the size of your files because continuous-tone images compress poorly in the GIF format.
If you get bad results with 256-color photos saved as BMP try saving them in the PNG format with your graphics program and then insert them directly into your Help+Manual project. This will ensure that they will not be converted in HTML-based output formats.
Smaller graphics take up much less space. Do you really need to include a full-size screenshot of the main program window? If the user is reading your help the program you are describing is already on the screen, and full-size screenshots also make the help difficult to navigate. A 75% or even 50% size version is normally plenty as a visual reference.
You can shrink the screenshot and add attractive callouts and other features with Impict, the full-featured screenshot editing program bundled with Help+Manual. You can also make screenshots at reduced sizes directly with Help+Manual's integrated screen capture utility or with the stand-alone TNT screen capture program.
Lots of full-color photos will also bloat your output files. If you do use them keep them small and if you are producing HTML Help or WebHelp make sure that your image conversion settings are set to Convert 256 colors to GIF and True Color to JPEG. While you are at it, experiment with the JPG compression setting in the same place; your pictures may be smaller and look just as good with a lower quality setting. A value between 70 and 80 is generally fine for most purposes and you can often get away with even higher compression settings.
Areas of an image that are a single color take up almost no space when the image is compressed and parts of an image that are not there take up no space. Quite often you can save space by cropping your images to remove unnecessary information, and this also focuses user attention on what is important.
You can also crop your screenshots with special effects like Ripped Paper Edge in the Screen Capture tool and the Impict screenshot editing program. This both saves space and makes your screenshots look better!
Help+Manual inserts JPG, PNG, SVG and GIF images directly, without converting or re-compressing them. They are exported to HTML-based output formats just as they are. However, they do have to be converted for some other formats. See SVG Graphics for some more information on this vector-based format.
You also can add vector-format Windows metafiles (WMF and EMF files) to your projects directly. When you publish your projects these files are converted to bitmaps and then the normal conversion and compression options set in your Image Conversion settings are applied.