This content has been marked as final. Show 10 replies
Hi KarenAa and welcome to our community
Can it be done? Sure, as long as you have the time and patience to sort through all the files. But the process will be long and painful and probably end up costing at least ten times in manpower (or more) than what you would save by just allowing the generation to create what it needs and publish the swarm to the server. Comparatively, disk space is quite cheap when compared with your time and sanity.
Yes, I know it's probably not what you hoped to hear, but it's the truth as I know it to be.
Sorry, forgot to add one thing. The ratio of "overhead" files to support different browsers may seem high and skewed at first. For example, a project that has 25 topics may produce 225 files. But most of those remain the same regardless of the number of topics. So as the project grows to perhaps 500 topics, the total may only be 700 or so.
Rick, the RoboHelp help topic "Output Files for WebHelp" states:
"...The output files located in !SSL!\WebHelp\whgdata are HTML data files used in Pure HTML versions of WebHelp."
"...The output files located in !SSL!\WebHelp\whxdata are XML data files used in Java applet and DHTML versions of WebHelp running on later browsers (such as Internet Explorer 5 and Netscape Navigator 6.)"
Reading that, it would seem like you could delete at least whdata (and maybe whgdata) if you knew your users were running on a late-model IE only. I haven't actually done any testing on this, but I have thought about it. Is there more to it than meets the eye?
One of my projects is 10.7 MB, and the supporting files are:
whdata: 210 KB
whgdata: 1,210 KB
whxdata: 265 KB
Unless disk space is an issue, why worry about it?
And, have you ever wondered where the TOC (.hhc), Index (.hhk), and Glossary (,glo) info is kept in the output folder?
An important issue here is how you are setting up the dialog that deals with dhtml, html and java applets.
For example, if you generate without Java applet you'll need whdata and probably whxdata too. If you're generating pure html, even IE needs whgdata, and the other folders probably won't take up much space.
I'm reminded of some general principles that you may disregard at some risk:
1. Don't throw data away unnecessarily. The information could turn out to be useful, if not essential.
2. Make no assumptions about what the user will have installed on his or her PC. You can state that a specific version of a third party application or utility is required, but that won't get you off the hook when a client has a problem. They don't like to hear that you told them so. Even when the outside app is a free download.
*Sigh.* Yes, this worry is why I didn't delete the files before and probably won't in the future. That said, we're probably in a better position than most to consider it, as our clients have to operate in a rather restrictive environment. We don't generate pure html, and our application requires the newer IE. Our help is large. I wouldn't quibble a bit about the extra space if it were one or two projects, but a quick search shows that the wh files take up more than 300 MB. (Burp!) It would be nice if we could carve a large hunk out of that figure.
But I won't hold my breath.
So what's the problem with 300 MB here and there, in the context of terrabyte-sized volume?
I know Harvey was teasing but he does have a point. We are talking about less than half a CD here and into freebie memory stick territory.
Against the risk of breaking the help in some way that nobody can foresee until months later when this is forgotten, it doesn't seem worth even a vague thought. I take your point about what is in the help but there's always the risk that it was either never 100% correct or that program changes have rendered it incorrect. Such things happen in any organisation. New developers, new authors etc, I'm with MergeThis. It isn't worth the risk for the saving.
Because this subject crops up from time to time, I have just added a Snippet on my site. I refer to RoboWizard's page on this subject.
I've also added one reason why I think sometimes authors get pushed into looking at reducing output size. Namely:
Sometimes the pressure to reduce the size of the help comes not so much from considerations of the server on which it will ultimately reside but on the media that will be used to deliver the application and the help. It can be as simple as trying not to go from CD to DVD or not going to one more CD or DVD than is already needed. Ask the developers if all their folders and files are needed. When they look horrified that anyone could be so stupid as to ask, just smile back.