I wrote:
CVS is used to provide & control access to the files.
We also have the option to use zope to provide distributed access. This would be a good public demonstration of zope power. Some advantages of using CVS: - most powerful version control & logging of who did what - widespread & standard - excellent tools like cvsweb ready-made - supports working offline & merging with master copy Some advantages of using zope: - demonstrates zope to potential users - simple through-the-web authoring - CVS is too tricky for many - advanced functionality can be added if needed - learning exercise I have been debating this zope/cvs issue myself, in regard to an intranet which I maintain which currently uses CVS to allow distributed authoring. Switching to zope will allow me to provide an easier interface & more features, but I don't want to give up the advantages of CVS either. I haven't thought of a good way to combine the two. Question: has the latest webDAV support resolved this dilemma, or will it soon ? Regards, -Simon
Simon Michael wrote:
I wrote:
CVS is used to provide & control access to the files.
We also have the option to use zope to provide distributed access. This would be a good public demonstration of zope power.
Some advantages of using CVS: - most powerful version control & logging of who did what - widespread & standard - excellent tools like cvsweb ready-made - supports working offline & merging with master copy
Some advantages of using zope: - demonstrates zope to potential users - simple through-the-web authoring - CVS is too tricky for many - advanced functionality can be added if needed - learning exercise
IMO, a major disadvantage of CVS is that it is mostly "file" oriented. In many (most) cases, multiple files evolve together. It isn't correct to use a particular version file without using compatible versions of other files. CVS has a feature to tag collections of files to represent a configuration of files with compatible versions, but you don't get to provide meta-data for tags and tags are an advanced feature that seems somewhat "tacked on". Zope, on the other hand, is transaction oriented. Zope tracks meta-data for transactions, not individual objects. This is a much saner approach, IMO.
I have been debating this zope/cvs issue myself, in regard to an intranet which I maintain which currently uses CVS to allow distributed authoring. Switching to zope will allow me to provide an easier interface & more features, but I don't want to give up the advantages of CVS either. I haven't thought of a good way to combine the two.
Question: has the latest webDAV support resolved this dilemma, or will it soon ?
It will impact this issue eventually. I can't say how soon. Jim -- Jim Fulton mailto:jim@digicool.com Python Powered! Technical Director (888) 344-4332 http://www.python.org Digital Creations http://www.digicool.com http://www.zope.org Under US Code Title 47, Sec.227(b)(1)(C), Sec.227(a)(2)(B) This email address may not be added to any commercial mail list with out my permission. Violation of my privacy with advertising or SPAM will result in a suit for a MINIMUM of $500 damages/incident, $1500 for repeats.
Jim Fulton Wrote:
IMO, a major disadvantage of CVS is that it is mostly "file" oriented. In many (most) cases, multiple files evolve together. It isn't correct to use a particular version file without using compatible versions of other files. CVS has a feature to tag collections of files to represent a configuration of files with compatible versions, but you don't get to provide meta-data for tags and tags are an advanced feature that seems somewhat "tacked on".
Zope, on the other hand, is transaction oriented. Zope tracks meta-data for transactions, not individual objects. This is a much saner approach, IMO.
I think that the basic underpinnings for this sort of development version control are present in Zope (the transactions), on feature I'd like to see in the Undo screen is some way to get at more of the meta data for a particular transaction - currently, the Undo screen shows you the method called, who called it, and when. However, usually, when I need to undo something, I'm thinking "I want to revert this DTMLMethod/Image/File to what is was..." i.e. I think about the object I've acted on, not how I acted on it. Sometimes that's encoded in the method URL, sometimes not. So, I guess if the transaction provided a view page that listed the Objects it modified, that'd be a big help. Hmm, I suppose I should go enter a wishlist bug... To continue the CVS comparison, the _next_ step is some form of diff - not only 'what Object changed' but 'how did it change?' Ross -- Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> NSBRI Research Scientist/Programmer Computer and Information Technology Institute Rice University, 6100 S. Main St., Houston, TX 77005
On Wed, 19 May 1999, Ross J. Reedstrom wrote:
To continue the CVS comparison, the _next_ step is some form of diff - not only 'what Object changed' but 'how did it change?'
There's another point. CVS allows you to jump to any point in time, and allows branches. Transactions let you jump back only, can't jump forward again, no branches. How about letting Versions be applied without closing the Version, and allowing a Version to be generated from a folder in the ZODB at a point in time. (I.e. makes the Version a copy of, or view into, the contents of a folder at a given instant.)
Ross
-- Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> NSBRI Research Scientist/Programmer Computer and Information Technology Institute Rice University, 6100 S. Main St., Houston, TX 77005
_______________________________________________ ZDP maillist - ZDP@zope.org http://www.zope.org/mailman/listinfo/zdp
-- Howard Clinton Shaw III - Grum St. Thomas High School #include "disclaimer.h"
On Wed, 19 May 1999 09:18:52 -0400, Jim Fulton <jim@digicool.com> said: [stuff about CVS...] Jim> Zope, on the other hand, is transaction oriented. Zope tracks Jim> meta-data for transactions, not individual objects. This is a much Jim> saner approach, IMO. Yes, one thing that Smalltalk and more recently, Java Beans, have taught us is that the file and directory structure was really just an artifact of the editing process. (Die-hard C and Makefile coders probably will disagree) Abstracting away from the filesystem or leaving it behind completely allows you much greater freedom. For example see a Smalltalk environment like Visual Works or Squeak, or for Java, IBM Visual Age for Java. The reason I'm kind of chiming in here, is that as a total Zope/Python newbie, it's kind of consfusing as a developer how Zope relates to my filesystem. Granted, the idea of a URL representing an object, method and parameters is very intuitive and appealing. But from a development point of view, it's it little rough going figuring out the alternatives to just typing DTML in an HTML form. Import and Export only can use pickled python objects right? In Visual Age you can import/export source code, compiled code, or archived code. In Zope the other methods add more complexity: FTP support is rough or broken, WebDAV is coming soon, Netscape PUT is rumored to work, not that great. I have no idea how all these ways of editing zope objects map into the "transaction oriented" model which you speak of. Alex Rice | alrice@swcp.com | http://www.swcp.com/~alrice Current Location: N. Rio Grande Bioregion, Southwestern USA
participants (5)
-
Alex Rice -
Howard Clinton Shaw III -
Jim Fulton -
reedstrmļ¼ wallace.ece.rice.edu -
Simon Michael