Why do XML editors suck so much?

Thursday, 2 October 2003

I’m seriously sick of using programs that call themselves “XML editors” because they colourize markup. I’m talking about XML Spy, Oxygen, BBEdit, and thousands of lesser programs. All of them are just glorified text editors - they still operate on the level of characters, not information items.

This is what I want to see:

Element selection - the primary selection mechanism should be per-element, not per-character. I want soft boundries on each element - maybe even go so far that if I want to select more, I should have to do something special, like hold a modifier key down.
Select-and-wrap - this is so basic I can’t believe they’re not all doing it. Select some content, right-click (or do whatever it is your platform does to get contextual menus) and allow me to wrap that content in an element of my choice.
Select-and-escape - Same as above, but do entity escaping, based on the document’s DTD.
Select-and-CDATA - Same as above, but wrap the selection in CDATA - intelligently.
Schema-aware editing - I’ve seen half-hearted attempts at this, but it’s usually too confusing to be useful. It needs to be live, contextual and unintrusive. Maybe squiggly lines?
Attribute editing - Some UI genius needs to come up with an intuitive, non-intrusive way to edit attributes. I shouldn’t have to position my cursor between the quotes, and I certainly shouldn’t have to worry about what style of quotes I’m using.
Non-continguous selection - Nisus Writer (a seriously good editor in its day) and Word do this - why can’t everybody else? Makes the select-and-* stuff a lot more powerful…
XInclude, XSLT, XML Encryption, XML Signature - all of these things should just happen if you ask for them. Bonus points for allowing people to mark sections of the document as XIncluded or encrypted, but still showing them the plaintext, so they can transparently edit and then save with the appropriate transforms.
Search-and-replace - Anybody think that maybe wildcards and regex isn’t the best way to find content in markup? How about XPath? How about “find me element X with attribute Y? And make it namespace-aware, please. Not just some poxy XPath navigator widget either - I want it to equal or better every other find UI in existence.
Not emacs - I don’t need a new religion, just an editor.

What else am I missing? Or is there already a perfect XML editor out there of which I’m blissfully unaware?

11 Comments

Steven Noels said:

While they won’t address all your issues, did you take a look at Pollo and XXE?

Friday, October 3 2003 at 2:20 AM

Mark Derricutt said:

I see XXE was mentioned above, this is by far one of the best XML “document editors” I’ve used. It still has a few oddities that set it apart as an XML editor, than than a document editor. But on the whole, its damn fine. And I find once you get used to using some of the key-strokes instead of the menus, things start running along nicely.

Friday, October 3 2003 at 3:33 AM

Danny said:

+1.

You could try pinging the LazyWeb, sometimes works.

btw, your Google Ads suggests ‘XML for Word’ and ‘Easy XML Editor’. Would you recommend either of these? (you don’t need to answer that ;-)

Friday, October 3 2003 at 6:04 AM

Doug Ransom said:

“Not emacs - I don’t need a new religion, just an editor. “

Sounds like you already have a religion. You just specified everything emacs can do, then repudiated its use because Jesus might not use emacs.

If you at all suffer from wrist pain, emacs is a blessing becuase the pointing device is not required. Emacs is hard to learn, easy to use. What does religion have to do with it?

Friday, October 3 2003 at 7:04 AM

Bill de hÓra said:

“Not emacs - I don’t need a new religion, just an editor. “

Oh well. You’ll have to go without James Clark’s nxml them, which is fast leaving all other XML editors in the dust. Pity :)

Saturday, October 4 2003 at 8:29 AM

artemis said:

“The primary selection mechanism should be per-element”

EditiX has this feature : http://www.editix.com

Monday, April 25 2005 at 12:02 PM

Craig Buchek said:

You should take a look at Arbortext (http://www.ptc.com/appserver/mkt/products/home.jsp?k=3591) and Morphon (http://www.morphon.com/xmleditor/index.shtml). They are WYSIWYG editors, that allow you to work with elements naturally. The WYSIWYG is more semantic-oriented than appearance-oriented, just like you’d want. I believe they both use CSS. I have not looked at either for several years, but they were quite good about 6 years ago when I was looking into them.

Wednesday, July 5 2006 at 8:32 AM

John Cowan said:

XED hasn’t been updated lately, but I rather like it. It’s a structure editor masquerading as a text editor. The only real downside is that because it insists on maintaining well-formedness at all times, you can’t even start editing an existing document until it is already well-formed, so if there’s a botched or missing tag you have to fix it in a text editor first.

Wednesday, November 1 2006 at 7:10 AM

Adrian said:

jEdit with the xml plugin covers a lot of this (tag matching, non-continguous selection, schema aware editing etc).

http://www.jedit.org

Wednesday, November 1 2006 at 12:39 PM

testor said:

Have your tried editix, this is for me a nice xml editor, I used it almost every days and I have exactly what I need. May be there’s something better, but this one is quite good for my usage (XSL and XML) http://www.editix.com

Saturday, March 24 2007 at 8:44 AM

seth park said:

You should take another look at oXygen.

They must have read your post…

Wednesday, November 7 2007 at 4:44 AM

mark nottingham

other XML posts