Unisys did market the Server Licencing as an 'alternative choice to better serve our customers' - instead of licensing the software that creates the gif's. However, as lots of Unisys stories suggest, the LZW patent helps feeding the poor families of a whole department of lawyers, and they don't take it easy once they are after you. I stumbled over an very interesting story recently: http://www.serverobjects.com/lzw.html However, Unisys shouldn't be legally entitled to go after per-server fees if the GIF's were created by legally purchased software (but it moght be worth a try). Unisys typically seems to include special statements for eval versions (i.a.w if the eval version can write GIF's, you might assume they are licensed) Regarding your 'burn all gifs' question: checking your server logs is probably the best to find out ;-) I see only some minor problems: a) Transparent Images - you might be using quite a few of them, right? I have no actual experience, but most Browsers are said to have problems with this. b) Your cute little animated emoticons... ---- Another issue are the images provided together with the articles. I'm not a lawyer, but as even sceptical human sense suggests: If you generate the 'standard' gif's with a 'legal' program, and enforce PNG or JPG for the images submitted together with articles, you should be on the safe side. Personally, from time to time I do work on a Win 3.11 system (It's in a lab, and Win9x+ has to much problems with real-time data acquisition without expensive extra hardware) - with just an Netscape 3. But it can be only good for science if I can't check CodeGuru from there... The BurnAllGIF's site (http://burnallgifs.org/) someone else mentioned in the conversation is up and running - and links to quite some GIF-Free sites. Good luck, and if you get in trouble with Unisys, you can rely on a huge developer community backing you. (Not that a lawyer is interested in this...) Peter

see http://www.4guysfromrolla.com/humor/images/Microsoft.jpg btw. the kid in the lower left corner would begates then, right

Uh... Is there anything that prevents 'people' from ripping my code? How good does the decompiled stuff look - how much of the original is retained? e.g. All internal function names & arguments? decompiling optimized C++ code creates rather obfuscated 'sources'. But if you get 'readable' sources back - woudln't this drive most companies away from developing for .NET? At least, drive them away from using .NET for desktop applications. (I still doubt the Internet is ALL the future

There is an alternative approach... Putting *Required* Processing Power into the Dongle, that is, not only some 'key verification', but something that is intrinsic to your app's function. A crack would require to re-invent the section of code placed into the dongle. Admittedly, this is only feasible for applications in the $1K+ range, and there are no development costs included. And it might cause a shift in the crackers profile, but it's worth the idea, isn't?

Maybe Bill likes SeaWorld, or maybe the beaches down that way are particularly fine :P