Fog Creek Software
Discussion Board




Knowledge Base
Documentation
Terry's Tips
Darren's Tips

Bug with no Workaround

I posted to the discussion group a week ago about a parse error I was experiencing when a CityDesk article with text copied from a Microsoft Word document and pasted into the article was served over HTTP. I still unfortunately have not resolved this problem and have not received any feedback from Fog Creek. I am currently evaluating this software for my company with the hope of purchasing several professional licenses however; my current workaround of linking to Word documents from articles is less than ideal.

I took the time to distil the problem down to a single line and emailed two HTML files (one working and one not) to support at CityDesk. One would imagine that the free testing would be appreciated, especially since this is clearly a bug in the CityDesk software. I was informed that CityDesk doesn’t provide technical support via email (unless of course I’m willing to pay extra for it). Although on the whole I’m impressed with CityDesk, I’m frustrated that Fog Creek hasn’t provided any support much less even acknowledged the problem.

Andrew
Friday, April 19, 2002

I've posted a similar problem recently, and although I got many answers (sometimes very useful) from the CD community, the FogCreek Team didn't answer specifically to this question.

It would have been nice from them to simply say in the board:
"we know the problem and are working on it - will be fixed in the next version".

Even if the solution comes a bit late, it would reassure potential customers that a solution will be found. Otherwise, we could fear that the problem will remain  with future releases, and this is a kind of "stupid thing" that can change a general positive opinion about the application into a non purchase decision.

That said, we can't say at all that the FogCreek support to people who have tested the product has been bad, especially during the beta test. I suppose that for a small company, e-mail support may quickly become an unbearable burden. Could they keep the application prize low if they offer a free wide support ? Don't know, but I think they have to keep CD prize low, and perhaps lower it again for the future versions if they want to sell a lot.

For me, I prefer a "low prize / paid support" compromise if the "low prize / free support" equation is economically impossible.

Perhaps support options should be more detailed on their website.

Vincent Bénard
Friday, April 19, 2002

Andrew,

The error you are talking about is not a bug in CityDesk.  Its a problem with the HTML code that MS Word generated and that you pasted into an article.  You said you get the "Parse error" when you view the page with IE, not with CityDesk.  CityDesk is not introducing the offending HTML.

Unfortunately, you are correct that we do not offer personalized technical support for CityDesk (email or other) for free.  But I hardly think its fair for me to say "It's not a CityDesk bug" and not tell you why ... so I looked at the files you sent us.

You sent us two files, one which you said worked fine and one which you said did not.  I couldn't repro your parse error in IE 6, but it seems logical to me that if one file is OK and one isn't, then the error is coming from the differences between the two files.  So I just opened them up and looked at them.  Doing a diff on the two files (one broken and one fine) shows that there is a line of HTML put there by Word that says:
<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />
If I were to make an educated guess, this is most likely your problem.  I don't know what it is, or why Word put it there, but I guarantee you that CityDesk isn't the culprit here.

I apologize for your frustration and I hope that this message helps you.

Michael H. Pryor
Friday, April 19, 2002

Michael,

Unless Microsoft changes office to be Citydesk compliant ;-) , the fact that the bug doesn't come from CD but from office is secondary. We don't search responsibilities but solutions to a very serious issue.

CD has to provide (necessarily) a way to clean up "cuts and pastes" coming from office, since office users won't have neither a way nor any interest to produce "cleaned" documents in a normal organization.

I've thought to a "setable parser" in a previous mail. I always think it's absolutely necessary. Currently the word parsing works a "nothing or all" way which is not satisficing.

The web site administrator could have an option that allows him to setup what must be kept, what must be thrown, and what must be changed and how from the original office formatting, and then set it as a default rule when a "simple user" pastes word or excel data.

Vincent Benard
Friday, April 19, 2002

There is a somewhat amusing tool called demoroniser http://www.fourmilab.ch/webtools/demoroniser/ that corrects Microsoft's HTML.

In the long run we'd love to be able to support code that cleans up the bad Microsoft "HTML" that comes from programs like Word but this is a major undertaking. In the meantime we do offer a limited "Paste without Formatting" feature which strips all formatting.

Joel Spolsky
Friday, April 19, 2002

Michael,

Unless Microsoft changes office to be Citydesk compliant ;-) , the fact that the bug doesn't come from CD but from office is secondary. We don't search responsibilities but solutions to a very serious issue. If office doesn't provide good code, CD has to bring a fix, otherwise who will use a software if he knows that office compatibility may suffer some problems with some office docs ?

CD has to provide (necessarily) a way to clean up "cuts and pastes" coming from office, since office users won't have neither a way nor any interest to produce "cleaned" documents in a normal organization.

I've thought to a "setable parser" in a previous mail. I always think it could help. Currently the word parsing works a "nothing or all" way which is not satisficing.

The web site administrator could have an option that allows him to setup what must be kept, what must be thrown, and what must be changed (and how) from the original office formatting, and then set it as a default rule when a "simple user" pastes word docs or excel tables. This parser could do the same job as dreamweaver "office cleanup" does, but perhaps with more flexibility.

a "standard parser" could be (non exhaustive)

keep: bulleted and ordered lists, block indent, bold and italic text...
change: doc fonts and styles to style sheet main font, underlined text to bold, simple line jump to <br>, dual line jump to </p><p>, more than two line jumps to </p>&nbsp;<br><p>, ... special chars by their &... counterparts,
throw away: all the <p style=...> <span class="...> garbage brought by office.

I guess it's not simple to implement, but you'll surely think to  smarter solutions for eliminating office cut and paste issues.

Vincent Benard
Friday, April 19, 2002

oops, my message has been posted twice, sorry for the noise.

Vincent Benard
Friday, April 19, 2002

When you say you are copying the Word document into CityDesk do you mean that you are saving the Word document and then importing the HTML page into CityDesk?

If this is the case, you may want to save the Word document as "Web Page, Filtered (*.htm, *.html)". I am using Word XP and I don't know if this option is available in other versions of Word.

What is does is remove all MS Office specific HTML from the document.

What version of Word and what browser are you using? I don't seem to have a problem with either generated Word file in IE5.5. And if I'm off the mark on what the problem is please let me know so I can try play with it. Thanks.

Stu
Friday, April 19, 2002

OK... I did try a bit more...

I created a web page in Word. Copied part of it and pasted it into a CityDesk article. I do notice the extra HTML blocks but it still comes up ok in IE.

Do you edit the text further in CityDesk or do you just publish it right after pasting? Could the HTML be getting mangled by the WYSIWYG editor?

Stu
Friday, April 19, 2002

I would have to disagree that CD has to solve this problem. Bugs have to be solved by the responsible party, in this case Microsoft. It is in fact impossible to solve this outside MS, as you don't know what exactly is wrong. You can just look at some output, and make a guess. That leads to kludges which are themselves likely to lead to further problems, e.g. when it is fixed by MS.

Just send a lawyer to MS and tell them to fix it. No CD development budget should be spend on fixing bugs in other programs

Stephan Eggermont
Saturday, August 03, 2002

*  Recent Topics

*  Fog Creek Home