[Q] My own editor?
Viva de el Zopistas, new to Zope, I have a few questions which I couldn't find the answer for. How can I export a DTML method file, so that I can add it again? Background: I want to use my own editor, as the webbrowser's text window usually is a real bad editor. (Of course I know, if the world would be more modular, I could just call my editor from the webbrowser. Most people are stuck with Mosaic-derived webbrowser (netscape/IE/Mozilla).) Another question: Would it not be advantagous only recreate a page, when the source objects change and trigger that creation? Most pages could just be cached this way. Or does Zope that already? Please cc: me answers, because I am not on the Zope list (just to huge). Regards, Bernhard -- Research Assistant, Geog Dept UM-Milwaukee, USA. (www.uwm.edu/~bernhard) Free Software Projects and Consulting (intevation.net) Association for a Free Informational Infrastructure (ffii.org)
On Sat, Nov 27, 1999 at 09:54:00PM -0600, Bernhard Reiter wrote:
Viva de el Zopistas,
new to Zope, I have a few questions which I couldn't find the answer for.
How can I export a DTML method file, so that I can add it again? Background: I want to use my own editor, as the webbrowser's text window usually is a real bad editor. From the containing folder, you can use Import/Export, but that generates a hard-to-read XML file. Another approach might be to connect to your Zope server with an FTP program on port 8021 and download/upload the file. Yet a third approach is constantly copying and pasting to/from your text editor and the web browser.
(Of course I know, if the world would be more modular, I could just call my editor from the webbrowser. Most people are stuck with Mosaic-derived webbrowser (netscape/IE/Mozilla).)
AFAIK, Mozilla has no Mosaic (or Netscape 4.x..thank God!) code in it whatsoever. Also, Ender, the text widget, is supposed to be extremely good. Several groups, including wxPython, KDE, and GNOME are working on modular web browsers, maybe someday you'll be able to insert your text editor as a component on a web page?
Another question: Would it not be advantagous only recreate a page, when the source objects change and trigger that creation? Most pages could just be cached this way. Or does Zope that already?
It doesn't do it at the moment. As your DTML gets more complex, it would become more and more difficult for Zope to determine when/if it needs to be reexecuted. Since DTML runs quite fast anyway, you'd quickly reach a point where determining whether or not a DTML page needs to be updated takes more time than actually updating the page. If you have an object that is really expensive to refresh (like a database query or something), then perhaps you could write a ZClass to encapsulate it and program your own logic to figure out if it needs to be updated.
Please cc: me answers, because I am not on the Zope list (just to huge).
Here here! I've been subscribed since 11/10/1999 and I've got almost 1000 messages in my "zope" folder in mutt. -- Stephen Pitts smpitts@midsouth.rr.com webmaster - http://www.mschess.org
On Sat, Nov 27, 1999 at 10:48:35PM -0600, Stephen Pitts wrote:
On Sat, Nov 27, 1999 at 09:54:00PM -0600, Bernhard Reiter wrote:
How can I export a DTML method file, so that I can add it again? Background: I want to use my own editor, as the webbrowser's text window usually is a real bad editor.
From the containing folder, you can use Import/Export, but that generates a hard-to-read XML file. Another approach might be to connect to your Zope server with an FTP program on port 8021 and download/upload the file. Yet a third approach is constantly copying and pasting to/from your text editor and the web browser.
All unsatisfactory. Using Zope looses you a lot of advantages you already have in your working environment. Like the unix tools box (if you are using Unix, that is), including syntax highlighting editors, version control, special script tools like find. In that respect it is not, what I was hoping it is. The cure to that would require a lot of unpleasant work, I am afraid. The Object, Aquisitions and Security model of Zope is great so far, but loosing the the other already existing nice components, unless they are rebuild for Zope is a huge tradeoff.
(Of course I know, if the world would be more modular, I could just call my editor from the webbrowser. Most people are stuck with Mosaic-derived webbrowser (netscape/IE/Mozilla).) AFAIK, Mozilla has no Mosaic (or Netscape 4.x..thank God!) code in it whatsoever.
I should have been more specific. The user interface and user logic is modeled similiar.
Also, Ender, the text widget, is supposed to be extremely good. Several groups, including wxPython, KDE, and GNOME are working on modular web browsers, maybe someday you'll be able to insert your text editor as a component on a web page?
To plug my personal opinion in here. I don't believe that KDE or GNOME will really produce modular and I mean task modular produts. At least they haven't done it yet. The modularity they talk about is different from the one I talk about, they create bloatware and don't solve tasks. (So much for a rant, I guess I'm not in my best mood today.)
Another question: Would it not be advantagous only recreate a page, when the source objects change and trigger that creation? Most pages could just be cached this way. Or does Zope that already? It doesn't do it at the moment. Okay.
As your DTML gets more complex, it would become more and more difficult for Zope to determine when/if it needs to be reexecuted.
It is object orientated and the objects know each other. If they are bidirectionally linked, someone could create a queue for each page. Each queue gets a timer value (like 30 Seconds). As soon as one object, this page depends on is changed, it schedules the page for rebuilding. When the page is called, it gets rebuild, if the queue flag is set and it was more than 30 seconds ago. Doesn't seem too hard to me at first glance. <wink>
Please cc: me answers, because I am not on the Zope list (just to huge). Here here! I've been subscribed since 11/10/1999 and I've got almost 1000 messages in my "zope" folder in mutt. This cries for an increase in the number of zope mailinglists. Or better: a zope newsgroup. Someone should initiate the process. Bernhard -- Research Assistant, Geog Dept UM-Milwaukee, USA. (www.uwm.edu/~bernhard) Free Software Projects and Consulting (intevation.net) Association for a Free Informational Infrastructure (ffii.org)
Bernhard Reiter wrote:
On Sat, Nov 27, 1999 at 10:48:35PM -0600, Stephen Pitts wrote:
On Sat, Nov 27, 1999 at 09:54:00PM -0600, Bernhard Reiter wrote:
How can I export a DTML method file, so that I can add it again? Background: I want to use my own editor, as the webbrowser's text window usually is a real bad editor.
From the containing folder, you can use Import/Export, but that generates a hard-to-read XML file. Another approach might be to connect to your Zope server with an FTP program on port 8021 and download/upload the file. Yet a third approach is constantly copying and pasting to/from your text editor and the web browser.
All unsatisfactory. Using Zope looses you a lot of advantages you already have in your working environment. Like the unix tools box (if you are using Unix, that is), including syntax highlighting editors, version control, special script tools like find. In that respect it is not, what I was hoping it is.
<snip> There are a few initiatives in process which aim to "pretty up" the text entry interface, ZIE being one (http://www.zope.org/Members/johanc/ZIE - works only with IE 5). I'd heard of others (VisualZope?) but haven't tried any. Zope's portability does come with a price. Zope runs on just about every widely-used UNIX variant, as well as NT. Its management interface has been designed to work with most every browser (it uses JavaScript at most very, very sparingly). The web interface is the *only* real management interface to Zope at the moment. Syntax highlighting: To my knowledge, there are no WYSIWYG editors that understand DTML tags, although there are of course many that understand HTML. Many can work with "virtual" filesystems via FTP. Use one of these Your other issues are pretty well-addressed by the software itself: Version control: Click on the Undo tab. Find: Click on the Find tab. No, they're not CVS or UNIX 'find', but Zope is not UNIX. -- Chris McDonough Digital Creations, Inc. Zope - http://www.zope.org
Hi Chris, thanks for answering my beginner questions. On Sun, Nov 28, 1999 at 06:19:02PM -0500, Chris McDonough wrote:
There are a few initiatives in process which aim to "pretty up" the text entry interface, ZIE being one (http://www.zope.org/Members/johanc/ZIE - works only with IE 5). I'd heard of others (VisualZope?) but haven't tried any.
Hmm ZIE is not an option. A general method for this would be nice.
Zope's portability does come with a price. I understand that. But loosing the modular toolbox of general computer systems is not really nice.
Zope runs on just about every widely-used UNIX variant, as well as NT. Its management interface has been designed to work with most every browser (it uses JavaScript at most very, very sparingly). The web interface is the *only* real management interface to Zope at the moment.
True. I am wondering a bit: Zope is such a special and big application, it would most certainly benefit from a real Zope client. (like written in Python/tcl or wxPython you would get all of the target platforms.)
Syntax highlighting: To my knowledge, there are no WYSIWYG editors that understand DTML tags, although there are of course many that understand HTML. I guess the syntax rules will be easy to add. And I will be content with a normal highlighting editor, it need not to be WYSIWYG, because that doesn't work on the web anyway.
Your other issues are pretty well-addressed by the software itself: Version control: Click on the Undo tab. Find: Click on the Find tab. No, they're not CVS or UNIX 'find', but Zope is not UNIX. And not NT for that matter ether.
Did you ever look into Hyper-G, now called Hyper-Wave? It had a nice attempt on a hypertext authoring system. Zope needs to be an hypertext authoring system to a certain extend. I am still missing bidirectional links in the world wide web, I guess. ;-> Bernhard -- Research Assistant, Geog Dept UM-Milwaukee, USA. (www.uwm.edu/~bernhard) Free Software Projects and Consulting (intevation.net) Association for a Free Informational Infrastructure (ffii.org)
Bernhard Reiter wrote:
On Sat, Nov 27, 1999 at 10:48:35PM -0600, Stephen Pitts wrote:
On Sat, Nov 27, 1999 at 09:54:00PM -0600, Bernhard Reiter wrote:
How can I export a DTML method file, so that I can add it again? Background: I want to use my own editor, as the webbrowser's text window usually is a real bad editor.
You could possibly use Emacs and it's FTP-on-the-fly features. Since I'm a bonehead Emacsist, that's my way of doing it. Well, mostly I mumble and curse with the Text widget interface....
Another question: Would it not be advantagous only recreate a page, when the source objects change and trigger that creation? Most pages could just be cached this way. Or does Zope that already? ... As your DTML gets more complex, it would become more and more difficult for Zope to determine when/if it needs to be reexecuted.
It is object orientated and the objects know each other. If they are bidirectionally linked, someone could create a queue for each page. Each queue gets a timer value (like 30 Seconds). As soon as one object, this page depends on is changed, it schedules the page for rebuilding. When the page is called, it gets rebuild, if the queue flag is set and it was more than 30 seconds ago.
Doesn't seem too hard to me at first glance. <wink>
Well, upon first glance: Could be. Just keep in mind that you can have computed dependencies, like referencing documents based upon runtime evaluated criteria, and you'll see that the very problem of checking 'what depends on which' is actually evaluating all and everything. IMHO this is a dead ringer for a Turing problem which means as much as 'To know what it will do when run you have to run it and watch'. A 'solution' like any arbitrary timer just isn't one: Imagine some changes on "important" properties like 'price' and 'amount' and 'sum', and then having to wait for those changes to 'settle' until all associated 'rebuild timer chains' have been rattled... I wouldn't trust such a system farther than I possibly could spit it ;-) Just my 0.02 ($§*@ Damn Euro symbol!) -- MfG: Tom Riedl Linux Information Systems AG | E-Mail: Tom.Riedl@Linux-AG.de "Linux is our Business" | Tel. : 089/99 34 12-11 http://www.linux-ag.de/ | Fax : 089/99 34 12-99
On Mon, Nov 29, 1999 at 02:24:47PM +0100, Thomas Riedl wrote:
Bernhard Reiter wrote:
On Sat, Nov 27, 1999 at 10:48:35PM -0600, Stephen Pitts wrote:
On Sat, Nov 27, 1999 at 09:54:00PM -0600, Bernhard Reiter wrote: You could possibly use Emacs and it's FTP-on-the-fly features. I will try the ftp feature.
Another question: Would it not be advantagous only recreate a page, when the source objects change and trigger that creation?
Doesn't seem too hard to me at first glance. <wink>
Well, upon first glance: Could be. Just keep in mind that you can have computed dependencies, like referencing documents based upon runtime evaluated criteria,
A 'solution' like any arbitrary timer just isn't one: Imagine some changes on "important" properties like 'price' and 'amount' and 'sum', and then having to wait for those changes to 'settle' until all associated 'rebuild timer chains' have been rattled...
I guess you are right. It would only be interesting for more non-dynamic pages. But it might be a real performance boost on these. :) Bernhard -- Research Assistant, Geog Dept UM-Milwaukee, USA. (www.uwm.edu/~bernhard) Free Software Projects and Consulting (intevation.net) Association for a Free Informational Infrastructure (ffii.org)
participants (4)
-
Bernhard Reiter -
Chris McDonough -
Stephen Pitts -
Thomas Riedl