RoboHelp Server 6 and NLS issues?

There seems to be a wildfire brewing among customers of Adobe RoboHelp Server 6.

Many have upgraded only to find that the legacy Natural Language Search functionality has not been included. Synonym search is gone, too. Apparently, Adobe has simply not included the NLS DLL file with the software.

What’s odd is that the functionality apparently hasn’t been disabled in the UI or the help. From the developer’s perspective, the feature just isn’t working. In other words, the features “appear” to be there, but aren’t. Users aren’t aware of the disabled functionality until they try to use it. Even the software’s help topics on NLS are apparently still present.

I’ve already gotten some comments from some who have encountered this.

Adobe has responded with an article in the RH Developer Center, alerting users to the “important change.”

Is it just me, or does seem a bit disingenous to alert users to the decreased functionality via a web article, rather than by removing the functionality from the UI?

It’s kinda like buying Microsoft Word, writing a letter, clicking Print, and then finding out that Microsoft didn’t include Print functionality with this version. So you open the software’s help to make sure you’re doing it properly. Yep, no problem… you think. Looks like I’m doing everything right. Finally, in frustration, you go to the MS website, and find an article “alerting” you to the fact that they didn’t put it in this time. No doubt you’d ask, Why did they include the ability for me to click “print” if printing wasn’t possible? Why include the help topics detailing the proper printing procedure, then? Odd.

Does anyone have suggestions for those who are dealing with this issue?

14 Responses to “RoboHelp Server 6 and NLS issues?”


  • I can’t wait to hear more about this. As I responded back to someone on the HATT server, I’ve been considering a Flare switchover, combined with a server technology. Right now it’s not even a question which company is more innovative in my opinion, I’m just looking for that server side component for reporting and most importantly, a solid search method.

  • HATT post, republished here. Monkey, feel free to nuke this if it violates your TOS.

    At the bottom is the reply that Peter Grainge, one of the long-time HAT gurus sent me. I’m now checking out his search solution for more specific information towards solving my problem.

    Thanks Peter,

    Actually, what I’m looking for is more in the line of synonyms, not like an ‘Ask Jeeves’ compared to Google. The phrase like a question is often called true natural language search capability, but not a lot of users want to type in proper english just to get results. This is probably where John (or Adobe) was looking at people not using it. However

    I should be very specific in what the industry considers NLS because there’s clearly not an agreement to what is a standard. Natural Language search is more than a phrased request. I’ll give some examples but basically it comes down to morphology and synonym recognition.

    I’m looking at both dictionary based and fuzzy morphologies. Ideally, these will detect all grammatical forms for words used in a search query, allowing for shorter query input covering different grammatical forms for each query word. Also known as synonym recognition. So the basic synonym recognition is dictionary based, at least as far as I know.

    Here’s how the search engine industry looks at search. Well, it’s how the search industry looked at it back in 2002 when I was looking into the sector. It’s what I use to quantify ‘best bang for my buck’.

    Engine features:
    Keyword and natural language search
    Boolean operators
    Phrase search
    Fielded search
    Word stemming, word breaking
    Metadata search
    Proximity search
    Wildcard search

    There’s a lot more for this, but this should give everyone a concept on what to search for in Google if anyone’s interested.

    So the next question becomes, just as in any keystone of a knowledge management project, how good is good enough? Do we want a $2000 server that will smoke Google’s $30,000 offering, or do we just want a boolean operator search that may or may not find what people are looking for.

    RoboEngine had a ‘pretty good’ engine, particularly for the price point. It had a lot of design issues which caused everyone pain in both supporting and designing. Adobe’s decision to remove those issues is the most cost efficient way to handle the issue. However… with the lack of those superpowered features (well, back in 2001 they were awesome, with time they’ve become standard) are my clients happy with the results?

    No. They want the same superpowers at the same price point. A good case example would be consumers and pizza. Nobody pays full price for pizza, they get the one on sale, because everyone’s always having a sale…

    So it comes down to a couple oversights. I say oversights because I’m certain the people who I know who still work at Adobe would not intentionally mislead users. Benefit of the doubt applies, right? It’s thin, but I’m giving the benefit of the doubt to Adobe.

    What does upset me is the price point. If you’re offering an ‘upgrade’ which in truth downgrades the performance point of your server, common consumer theory says that the price point should reflect that. Second issue I have is that even though I’ve been reading the RoboHelp product manager’s blog, and recently that monkeypi site (http://monkeypi.net/?p=162) just opened a thread on it.

    I’ve seen a lot of bold talk about how features have been enhanced, this is truly getting to be irritating to separate the media flak from the truth. Fine, so I get the same thing I had within RoboHelp WebHelp (text-based search) but it tracks results? Cool, make it a cheaper add-on that truly represents the functions added. $2000 should buy me a FUNCTIONAL, fully supported version which includes innovation instead of 2001 technology in 2007.

    Instead we’re looking at 1995 technology within a 2007 application. It appears to be a patch upgrade. I can’t sell this and it makes me look fat wearing it. ;-)

    My pals within the RH dev group (and you know who you are) I still love ya but you cannot tell anyone with a straight face that this is progress. I know you all probably fought the good fight getting even the latest changes done in RH, but selling this as a market leader is going to get people fired. Yours or ours, we are going to lose face marketing a falsity. And stating that RH Engine is improved, and nobody uses Natural Language Search is borderline falsity. As I said before, I’m going to consider it an oversight because I know the people I worked with at eHelp five years ago wouldn’t sleep at night if they were responsible for the spin on this. Don’t make me think all the ethical people left, I know that can’t be true.

    What I want: let’s have some truth in the game first, and second, I still still still still still still still am looking for a solution. Otherwise I’m not going to make money on my proposals.

    Sincerely,

    Charles

    — In [email protected], “Peter Grainge” wrote:
    >
    > I suspect John is safely tucked up in bed right now but I believe from
    > exchanges with him he is referring to NLS in the sense of “where can I
    > buy a black Ford Explorer” rather than just popping in “black Ford
    > Explorer”. John can comment on that later.
    >
    > I wonder whether ZoomSearch would fit the bill for you. There’s a
    > topic on my site about it and the search there uses it. It’s very
    > configurable, easy to use, inexpensive and the support is excellent.
    >
    > http://www.grainge.org
    >
    > ________________________________________
    >
    > Peter Grainge

    Update: For $300, Peter’s solution appears to go well towards solving a lot of my issues for very little dough…

  • I have seen the information on your site about both RH and Flare, but has anyone looked at or mentioned how it compares to ec software’s Help & Manual? Reading up on H&M, it seems to be in the same class as the other 2, or am I missing something?

  • Well, in this string you’re missing the complete discussion about the search engine component.

    But… let me gaze into my crystal ball for a second and predict something based on the sock puppeting around here. (at least this puppet smartened up and didn’t put his web address onto his link)

    (clears throat)
    Adooobeeeee is going to buyyyyy a newwwwww tooool….. It’s going to solve… the problems with RoboHelp…. It’s going to be cheaper… stronger… better… faster…

    And a local search engine company will provide the backend.

    Tell me I’m wrong.

    You guys need to check the cluetrain manifesto (www.cluetrain.com) and stop insulting my intelligence with two links in two days in two threads about some obscure help authoring tool that has nothing to do with my valid and urgent need for an NLS search engine. Do I really really reallly have to go back to my 2002 NSL contacts and custom build something instead of buying one off the shelf?!?

    “Will nobody rid me of these meddlesome priests (product evangelists)?!?!?”

    Item: Nobody at Adobe has responded to my post on their product manager blog for RoboHelp. I’ll send it to MonkeyPi so he can decide whether or not it fits his forum but it deals with NLS.

    Item: Some clown posts twice in two days about the same product in two different threads. No doubt keeping his dream alive for his company specializing in either search or some other Adobe partnership soon to be announced. Hmm. Search engine XML based no doubt with a cheap RH interface. Sorry, that was innovative around 2004. I know, I thought of it and mentioned it to a Macromedia executive.

    Item: Nobody has jumped into the search fray, yet RJ cowboyed up enough in the last forum about RH to drop his phone number. That guy has balls, I personally dig him which is why Adobe will probably blame all this dissent on him. Once again, the cluetrain will pass through the station and not be caught.

    So that’s what you get for posting off topic with no contact info.

  • “So that’s what you get for posting off topic with no contact info. ”

    dude, dude yer missing the whole point, this site is cool about annominity :-]

  • ;-) Yep.. But I would at least like to find out more topic information. Anonymity is cool but spammers also enjoy anonymity and give no information back.

    NLS - topic at hand… but I could be harshing someone’s gig. I apologize at general if I’m too intent. I’ve got to find something to solve this search component for the right price.

  • In the same line, has anyone noticed that other functionality included in RoboHelp 5x is missing in version 6, namely, the ability to create and use relative links in the table of contents in merged WebHelp.

    This issue has made my life a living h311, as I have just finished building a documentation intranet using RH5, with merged WebHelp that has thousands of topics and links. Due to the lost functionality, I may have to start over.

    The story: Merged Webhelp was created with a main project and about 15 subprojects. One subproject contained the majority of the HTML documents (with PDF attachments) to meet our business requirement of maintaining one copy (and one only) of each document. Other subprojects linked back to those topics. Now, the subproject-to-subproject linking does still work, but relative links in the table of contents are no longer supported. And, I have LOTS of those.

    Guess what? I am looking for a new tool, and I have nearly a year’s worth of work down the drain.

    How often does a software release REMOVE functionality that you’ve had for years?

    Note: It’s not for lack of trying. Tech support has looked at the issue. Also, Peter Grainge has worked extensively on this issue and determined that some file errors that made it through without too many consequences in RH5 will not “pass” in RH6, so some problems were on my end during conversion. Subproject-to-subproject linking works. However, we validated that the relative links in the TOC would not work, no how, no way, however. For more information about using RH6, see Peter’s site at grainge.org.

  • I am rather staggered by Pro.Techwriter’s post as it tells only part of the story.

    I had worked with Pro.Techwriter on trying to find out what was going on when the problem was posted on the RoboHelp forums. What was being reported about what Support had said was at variance with my understanding and with what someone else in Support had said to me. I wanted to investigate to find out the true answer as well as to help Pro.Techwriter. After a number of posts, I was sent part of the project and was able to confirm for myself that a relative link in the TOC to an external topic, including a topic in another project in a webmerge setup, worked in X5 but not RH6.

    On 11 April I contacted Adobe saying that this issue was a real problem and needed to be addressed. On 19 April I advised Pro.Techwriter that a patch had been produced (just 8 days later) and that I had tried it on her project and it worked. By that time a decision had been made by Pro.Techwriter not to proceed with RoboHelp which is of course her right. However, to post on 27 April “we validated that the relative links in the TOC would not work, no how, no way” is, in my opinion, grossly unfair. At one point that was true but 8 days earlier I had advised a patch was available that did fix the problem on the part of the project I had.

    Pro.Techwriter was not free to post about the patch at that point but to give the impression that there is an unresolved issue is wrong. At the time Support were contacted, they could not say a patch was being worked on because that only started later.

    The patch also fixes the problem that has been reported about generating help when a project has merged cells in a table and a build expression is used. It is available from Adobe Technical Support and from my site.

  • Peter, you rock. I find myself following up on the HATT forums whenever I have time to give out tips on RH to the users also, and it’s great to see your post on this board.

    My question is:
    Any ideas about that previous issue I have had with the RH server capability?

  • Charles, I’m scratching my head a bit as I thought the response to look at ZoomSearch had met with your approval and it did most of what you wanted? It didn’t give you a true natual language search but I thought you indicated that what it gave was close enough?

    Feel free to offline me.

  • Sorry. That was a typo that changed the meaning - it should read any further ideas about that previous issue.

    Thank you, I liked the $300 solution but one of my clients didn’t so I’m back to the drawing board… Thought since you posted here you might have some further ideas.

  • I read it the way you describe but what threw me was why you would want further ideas, now I understand. Sorry but once I found ZoomSearch I stopped looking. I had seen some other stuff but nothing that was so easy to use and effective.

    Out of interest, what was it they didn’t like, surely not the price tag?

  • Hi Peter,

    Believe it or not, they wanted a better NLS, not just indexing and word stemming. They decided that the synonyms weren’t capable enough in this one for the rollout they were doing because it would take a bit more time down the line to customize it on an ongoing basis.

    Maybe they didn’t like the time estimate in how long it takes to maintain it, or something else they didn’t tell me. I think they really liked going to a one-stop solution. RH Server was complex however, and server support at Adobe wasn’t thrilling them either.

    If you have any great tips or a comparison between the RH Server and ZoomSearch I will quickly pass them onto my client!

  • I’d love to know what they eventually settle for!

    Synonyms could take some time to set up but but the creation of the word searching is as long as it takes ZoomSearch to trawl through all the topics, Word documents and PDFs. I ran it across a pretty large repository of documents and at most it was an hour or so including all the uploading etc.

    I wonder if they have a fixed idea of what they want and have not tried ZoomSearch on their users to see if, shock horror, it finds what the users are looking for.

    I don’t have a comparison to offer. I am sure you have seen MergeThis post on the RH forums and I know he is sold on ZoomSearch. Uses the categories that I haven’t played with. When it didn’t do something quite as he needed, he emailed the developers and they tweaked the product for him. Might be worth another post there.

Leave a Reply