Linking to ISAPI web application - effect on SEO

Appreciate the quick response.  Do you happen to know of any on-line article or such like that backs this up?  Not doubting what you say in the least but I have a customer who claims he has noticed a drop in ranking since adding the links and I would like to have something to back up my rebuttal.
Thanks Again.

To be more specific about this particular case, I would have to take a look at the app.  Also, I have not seen reference made to this specific configuration before.  However, having a good overview of how search engines work (Google FAQ - www.google.com/support/webmasters/ - and quality guidelines - www.google.com/support/webmasters/bin/answer.py?answer=35769), SEO fundamentals (www.seo-blog.com/tutorial.php), and SE ranking factors (www.seomoz.org/article/search-ranking-factors), you can see that there is nothing inherently problematic with the approach you have taken.

With additional reflection, the only potential problem I can think of would be an issue of a SE perceiving the app as multiple orphan pages (inbound, but not outbound links).  If a website has too many dead end pages, it can impact on the SE ranking to a reasonably minor degree.  If each of the multiple links to the app dynamically generate a different URL, Google (et.al.) could see those as multiple dead end pages.  However, the type of setup you are talking about is fairly conventional and isn't the sort of thing SE's would seek to penalize in any way.

Something I should have added in the first place, if you can exclude the app from being crawled using a robots.txt file (www.robotstxt.org/robotstxt.html) and/or a "nofollow" meta tag (www.robotstxt.org/meta.html), you can make clear that the target of the link is not meant to be indexed.

Links appreciated.  

Something I failed to mention was that the web application is (and must be for commercial reasons) hosted on a different domain.  Does this affect your answer in any way?  

To further clarify.  Any single page could have many links to the same dll applicaton, each calling the same program but passing different parameters.  So to use the robots approach, if I have www.existingsite.com that links to url on a different domain to run the web applcation (e.g. www.webappdomain.com/cgi-bin/webapp.dll?a=1&b=2 etc) then I could use a robots.txt in the format...

User-agent: *
Disallow: /http://www.webappdomain.com/

.. to exclude all links to that domain from any bot?

Links to another domain can not be managed with a robots.txt file, so that is out in this case.  Stick with the 'nofollow' tag.

Linking to a different domain does have different implications, though.  The quality of the sites you link to reflects on your site in the eyes of the search engines.  For example, linking to a 'bad neighborhood' (ie. link farm, banned site, etc.) will negatively impact on your site's ranking.  Unless webappdomain.com has issues along these lines, it shouldn't be a problem.

ok so robots is out but should I not be looking at www.robotstxt.org/faq/relnofollow.html instead of
www.robotstxt.org/meta.html ?

The META based tag on the source page would mean that (www.existingsite.com) would not be indexed (which would be disatrous) and the same could not be used on the target page (www.webappdomain.com) as the HTML is only dynamically generated when the dll is executed and so would be invisible.  I presume you mean something like this on the source page?

<a href="http://www.webappdomain.com/...." rel="nofollow">run application</a>

Many thanks

Yes, exactly.  I linked to the wrong page there.  You want the tag on the link, not on the page.  I'm glad you picked that up and thanks for pointing it out.

ok (and before you read on I apologise for the misinformation). One further clarification which I overlooked.  The links are also dynamically generated so instead of

<a href="www.webappdomain/...">run application</a>

we have

<script language="javascript" type="text/javascript">
document.write(buildlinkhtml(param1, param2, param3));
</script>

I could easily add the rel="nofollow" to the javascript function the returns the link html but as the link itself, as well as the page generated by the application, would be invisible, are we back to your original opinion that we have nothing to worry about?

The search engines try to extract links from javascript, flash, etc., but still aren't very good at it.  If you are using javascript links, they may not be picking up the link at all.  Nevertheless, if that is the approach you are going to take, you may want to try to include the nofollow anyway.

Something else to consider - if the website is primarily using javascript navigation, you can pretty much count on the search engines having trouble with it.  Overall, you should be looking for other SEO problems anyway if the site has been struggling and the navigation structure may be a good place to start.

ok, have added the nofollow to the js code to cover all bases.  

All other navigation on the site that we want to be indexed is standard <a> tags and so the site is performing ok in the rankings.  I am only planning to add this navigation method (as it is necessary for the links to be built dynamically) to call the web applciation and will leave all other navigation as is.  This takes us back to my original question of wanting to be sure that this additional linking method would be ranking neutral.

Yes, it should be.

ok, I'll go with that. Thanks for all your input.  Much appreciated.

Glad to be able to help.