Menu
▾
▴
warp-framework-devel — Discussion related to the development of the Warp(ed) framework
You can subscribe to this list here.
| 2001 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(277) |
Nov
(3) |
Dec
(15) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 |
Jan
(1) |
Feb
(5) |
Mar
(45) |
Apr
|
May
|
Jun
(1) |
Jul
(6) |
Aug
(2) |
Sep
(2) |
Oct
|
Nov
|
Dec
|
|
From: Myroslav O. <myr...@zo...> - 2002-09-02 22:59:38
|
Myroslav Opyr wrote: > I've come to very unpleasant situation, where I am unavare what to do=20 > and how to improve situation. There was a need of wf-based application=20 > transfer with zexp import/export. When Importing aplication amount of=20 > objects Catalog doubles. Objects on their original locations and=20 > objects on their new locations. Just Update Catalog does not help=20 > because of a lot of references and circular references in the=20 > application (Update ends in inability to find object with specific id,=20 > which is not in Catalog yet). Clear and then Find and Catalog suffers=20 > the same problem. update_catalog_reindex_batch is inefficient at all. > > What manual procedure of Catalog repair should look like?=20 Ok. I've just discovered that adding meta_type =3D=3D '<mt>' helps. I rep= eat=20 the procedure for all meta_types in the application and then do=20 update_catalog_reindex_batch several times to flush incorrect cached=20 data out of Catalog. I am not sure if everything is working properly=20 after such procedures, but it at least start working. The question is still valid... > How can I correct the code to avoid such problems. I do not see light=20 > in the end of the tunnel yet :( m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> |
|
From: Myroslav O. <myr...@zo...> - 2002-09-02 22:06:43
|
Hi, I've come to very unpleasant situation, where I am unavare what to do=20 and how to improve situation. There was a need of wf-based application=20 transfer with zexp import/export. When Importing aplication amount of=20 objects Catalog doubles. Objects on their original locations and objects=20 on their new locations. Just Update Catalog does not help because of a=20 lot of references and circular references in the application (Update=20 ends in inability to find object with specific id, which is not in=20 Catalog yet). Clear and then Find and Catalog suffers the same problem.=20 update_catalog_reindex_batch is inefficient at all. What manual procedure of Catalog repair should look like? How can I=20 correct the code to avoid such problems. I do not see light in the end=20 of the tunnel yet :( Regards, m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> |
|
From: Myroslav O. <myr...@zo...> - 2002-08-07 21:11:10
|
Hi, There is 'widgets' attribute defined in reference_factory and in=20 multiple_reference_factory. Can you explain what they were for? I=20 understand that there should be that types of widgets... I need popup=20 and other types of widgets (iframe, r/o href, select+href, etc.) now.=20 Can you explain design ideas you had in implementing them. This would=20 correct me when if I'm lost... Regards, m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> cell: +380 50.3174578 |
|
From: Myroslav O. <myr...@zo...> - 2002-08-01 23:42:14
|
Hi, Morten, I'm a bit surprised. I've tried to mimic the behaviour of Eric'=20 address_book and hide subclass similar as Eric did for 'Groups=20 container', but failed. My 'Types Container' is continuously appearing=20 in the places it should be hiddedn even when I've registered it with=20 visible=3D0. What it can be? m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> cell: +380 50.3174578 |
|
From: Myroslav O. <myr...@zo...> - 2002-07-19 01:00:36
|
Hi, Could you please take a look on modifications I've done in 'interra'=20 branch. The most dangerous is id->warp_id transition, it started but had=20 not finished yet. I tried to make it smooth enough. Others are bug-fixes=20 and usability/aesthetic improvements. Comments should be helpful. I'd like to get feedback and recommendations. Should I continue in my=20 trunk or I there is good idea to merge with thrunk? IMHO I have to=20 finish id->warp_id transition and write converters for existing=20 applications (probably make them run automagically upon detection of=20 warp_id absence). Only then merge with trunk should take place. m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> cell: +380 50.3174578 |
|
From: Morten W. P. <mo...@ni...> - 2002-07-14 12:54:09
|
Myroslav Opyr <myr...@zo...> said: > Morten W. Petersen wrote: > > >Myroslav Opyr <myr...@zo...> said: > > > >>Can you apply the patch to html_edit_properties_display_light.py? When > >>translating code from DTML to Python there was mistake: > >> > >> > >I thought you had access to the CVS tree? > > > > > Sorry ;) I didn't noticed. I'll report problems if any... > > What can I do with the code? What kind of commits you want me to > discuss? What kind of them I can do without your confirmation? > > Or maybe it would be better to have just separate branch for my code > modifications? A separate branch would be best. I've upgraded your status to project manager on the project, that means you can do pretty much everything you want. Have fun. :) -Morten |
|
From: Myroslav O. <myr...@zo...> - 2002-07-14 12:29:08
|
Morten W. Petersen wrote: >Myroslav Opyr <myr...@zo...> said: > >>Can you apply the patch to html_edit_properties_display_light.py? When=20 >>translating code from DTML to Python there was mistake: >> =20 >> >I thought you had access to the CVS tree? > =20 > Sorry ;) I didn't noticed. I'll report problems if any... What can I do with the code? What kind of commits you want me to=20 discuss? What kind of them I can do without your confirmation? Or maybe it would be better to have just separate branch for my code=20 modifications? Regards, m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> cell: +380 50.3174578 |
|
From: Morten W. P. <mo...@ni...> - 2002-07-14 12:23:49
|
Myroslav Opyr <myr...@zo...> said: > Hi, Morten, Hi Myroslav, > Can you apply the patch to html_edit_properties_display_light.py? When > translating code from DTML to Python there was mistake: I thought you had access to the CVS tree? -Morten |
|
From: Myroslav O. <myr...@zo...> - 2002-07-14 12:08:39
|
Hi, Morten, Can you give me a guiding light? As far as I understood wf.user want=20 some entity to provide get_home_folder() function returning the=20 homefolder of the user. And it is unlikely that user.py offers something=20 like that. Thus I've got an idea that it should be some special user=20 AuthenticatedUser descendant created by some special UserFolder isn't=20 it? What it was in your case. I just do not want to invent a wheel ;) As=20 far as I've seen that user has catalog (for his personal use: bookmarks,=20 etc.), notifications (messages to display), get_workspace_actions(), and=20 browsing_history folder containing 'Bookmark history entry' objects. And having seen get_messaging_client() reference in=20 address_book.email_address invoked with get_user_folder() and with=20 get_user().get_home_folder() makes me think that there should be some=20 special wf.user_folder. Is there any chance to have a look on it? Regards, m. --=20 Myroslav Opyr zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting e-mail: myr...@zo... <mailto:myr...@zo...> cell: +380 50.3174578 |
|
From: Myroslav O. <myr...@zo...> - 2002-07-14 09:59:24
|
Hi, Morten,
Can you apply the patch to html_edit_properties_display_light.py? When=20
translating code from DTML to Python there was mistake:
line 40
html +=3D property % {
'title':property_.title,
- 'widget':property_.widget(keywords['name_prefix'])
+ 'widget':property_.widget(name_prefix=3Dkeywords['na=
me_prefix'])
}
m.
--=20
Myroslav Opyr
zope.net.ua <http://zope.net.ua/> =9C Ukrainian Zope Hosting
e-mail: myr...@zo... <mailto:myr...@zo...>
cell: +380 50.3174578
|
|
From: Myroslav O. <My...@or...> - 2002-06-20 19:14:10
|
Myroslav Opyr wrote: > Morten W. Petersen wrote: > >> On Wed, Mar 20, 2002 at 12:25:18AM +0200, Myroslav Opyr wrote: >> >>> Morten W. Petersen wrote: >>> >>>> Why are friendly URLs essential? >>> >>> It is human-factor. >>> >>> 1. When I look at friendly url I can guess what is going on even when >>> I know nothing about the system. >>> 2. Statistical analysis (webalizer output) of application (for >>> usability improvements) is much easier if I see readable URLs. >>> 3. There is no need to use extra tools for crash analysis except logs >>> with friendly URLs. >>> 4. Other options I have no imagination now, but they exists. >> >> OK, I'd like to see a 'spike' (example code) for this, to see how it >> could be implemented. > > First I'd like to remove Zope ID dependency. > > To achieve this I'd add immutable warp_id (or wid to shorten) to each > object within WF application (within scope of id_generator). Upon > creation (or clonong) each WF object should get an Unique warp_id and > register itself in Catalog. Later all refernces within application > should be warp_id based. Catalog would be the engine which resolve all > such references. > > The described above change is not spike scope one. It is major system > modification. catalog.py and all get_catalog() related code should be > modified at least... Done. I have code that is working. warp_framework application got an wapr_id of 0 and all other objects have generated warp_id. Code was partly upgraded to use warp_id instead of id (not everywhere but all places were marked). > In the beginning I'd left ID generation as is but later it would > change to allow arbitrary id changes and object location within object > hierarchy (as soon as the object updates Catalog after such moves). It is working already. ID is in form of <meta_type>_<warp_id>. I'm coding patient register now so I'll make it overridable. m. |
|
From: Morten W. P. <mo...@th...> - 2002-03-26 22:41:31
|
On Tue, Mar 26, 2002 at 11:49:54PM +0200, Myroslav Opyr wrote: > Morten W. Petersen wrote: > > >Both the boxed subclasses and widget sheet schemes were 'front-runners' > >to the 'edit multiple objects on one page' concept. I agree, they're > >ugly, and they don't really do anything useful now.. > > > >I'd say remove them both. :) > > > Meaning? Do you propose to replace them with something better? Yes, replacing them with something better. The boxed subclasses concept could be useful, I'm not sure. > Concept: 'edit multiple objects on one page' > I'd extend it into 'multiple objects on the one page', it is not > necessary 'edit' it can be 'view' oe 'control'. > > Distant goal: > improve usability. > Immediate goal: > reduce number of mouseclicks to complete misc taks. > Means: > One of the ways is to represent as much information on one page as > possible (still avoiding scrolling). > > I see that it is contrast to my previous mail in this thread mentioning > "bloated" interface. It is contrast but it is not conflict. interface > can be condensed but clear and easy to follow. I agree. > P.S. What about other open questions? I am stuck unless I get answers... I'm very busy packing, doing some final errands & lunches (going to Norway tomorrow) so I don't have time to answer them straight away. I'll be taking a long Easter break and talking to prospects so it might be a while before they get answered, just so you know. Regards, Morten |
|
From: Myroslav O. <My...@or...> - 2002-03-26 22:35:50
|
Morten W. Petersen wrote:
>On Sat, Mar 23, 2002 at 04:04:47PM +0200, Myroslav Opyr wrote:
>
>>I've discovered that it is possible to change the GUi of an object by
>>adding implements_boxed_subclasses and implements_widget_sheets
>>functions. IMHO the names of the functions and hackish way they are
>>hooked into framework is not good solution. I have some ideas regarding
>>the GUI display screens but it is subject of another e-mail.
>>
>>Amount of information available both in boxed_subclasses and
>>widget_sheets views is enormous. "Weight" of the page is big too, thus
>>loads and renders much slower. But even this is not an issue. The issue
>>is usablity. Something is wrong with them. Screens are bloated with
>>control aread, buttons, etc. Users (including me ;) get lost in these
>>screens.
>>
>Both the boxed subclasses and widget sheet schemes were 'front-runners'
>to the 'edit multiple objects on one page' concept. I agree, they're
>ugly, and they don't really do anything useful now..
>
>I'd say remove them both. :)
>
Meaning? Do you propose to replace them with something better?
Concept: 'edit multiple objects on one page'
I'd extend it into 'multiple objects on the one page', it is not
necessary 'edit' it can be 'view' oe 'control'.
Distant goal:
improve usability.
Immediate goal:
reduce number of mouseclicks to complete misc taks.
Means:
One of the ways is to represent as much information on one page as
possible (still avoiding scrolling).
I see that it is contrast to my previous mail in this thread mentioning
"bloated" interface. It is contrast but it is not conflict. interface
can be condensed but clear and easy to follow.
Regards,
Myroslav
P.S. What about other open questions? I am stuck unless I get answers...
|
|
From: Morten W. P. <mo...@th...> - 2002-03-26 13:27:22
|
On Sat, Mar 23, 2002 at 04:04:47PM +0200, Myroslav Opyr wrote: > I've discovered that it is possible to change the GUi of an object by > adding implements_boxed_subclasses and implements_widget_sheets > functions. IMHO the names of the functions and hackish way they are > hooked into framework is not good solution. I have some ideas regarding > the GUI display screens but it is subject of another e-mail. > > Amount of information available both in boxed_subclasses and > widget_sheets views is enormous. "Weight" of the page is big too, thus > loads and renders much slower. But even this is not an issue. The issue > is usablity. Something is wrong with them. Screens are bloated with > control aread, buttons, etc. Users (including me ;) get lost in these > screens. Both the boxed subclasses and widget sheet schemes were 'front-runners' to the 'edit multiple objects on one page' concept. I agree, they're ugly, and they don't really do anything useful now.. I'd say remove them both. :) Cheers, Morten |
|
From: Myroslav O. <My...@or...> - 2002-03-26 10:08:48
|
Myroslav Opyr wrote: > Morten W. Petersen wrote: > >> On Tue, Mar 19, 2002 at 11:06:03PM +0200, Myroslav Opyr wrote: >> >>> Today I got strange (for me at least) compatibility problems. I >>> tried to move warp-framework based application from one Zope >>> instance to another one. Additionally it was platform change, I >>> moved from Win32 to FreeBSD4. I used zexp for "migration". I partly >>> solved the problem with * * thread_safe_counter (recreated it's file >>> in var dir) and having updated Catalog. But I got stuck with >>> problems appearing when instance of any object is created. It just >>> refuses to render the page (object_manager_index containing the >>> object). Object is creatred and deleting it from ZMI solves the >>> problem. >> >> What if you create a new instance on FreeBSD4 and add data to it, does >> that work? > > Yes it works. I've created another instance. None of the symptoms > above... I'll continue experiments to find what is causing the problem. > > Myroslav Scenario: Export WF-based application from Win32 into zexp Import zexp on FreeBSD (same Zope 2.4.3) Update Catalog -> working Rename application -> not working Update Catalog -> working again Adding an object sometimes work sometimes causes object_manager to fail render page at all. Any ideas? I continue investigation the case because transportation is really necessary. Myroslav |
|
From: Myroslav O. <My...@or...> - 2002-03-24 01:41:30
|
Hi,
I was wondering if we follow "Plone style guide"? It states this:
* Plone has two types of buttons
o White means "this button will do action depending on data
youentered/items you have checked"
Examples: Standard submit buttons, the rename/copy/delete
buttons
Name: "context"
o Blue button means "go to a page where you can alter stuff"
Examples: Local Role button - this brings up a new page
where you can enter info and save it.
The save button in itself is white, as it does an action
based on entered data.
Name: "standalone"
I am just impressed with the look and feel of CMFPlone 0.9.9. Now I see
where it is possible to move... What do you think about updating to
current version of Plone?
Regards,
Myroslav
|
|
From: Myroslav O. <My...@or...> - 2002-03-23 14:06:07
|
Hi, I've discovered that it is possible to change the GUi of an object by adding implements_boxed_subclasses and implements_widget_sheets functions. IMHO the names of the functions and hackish way they are hooked into framework is not good solution. I have some ideas regarding the GUI display screens but it is subject of another e-mail. Amount of information available both in boxed_subclasses and widget_sheets views is enormous. "Weight" of the page is big too, thus loads and renders much slower. But even this is not an issue. The issue is usablity. Something is wrong with them. Screens are bloated with control aread, buttons, etc. Users (including me ;) get lost in these screens. Sorry for critique only. I am not expoing proposals untill I get feedback to this critique. I'd like to know what stopped Morten from doing things in this or that way. Then I would avoid such problems from the very beginning. Regards, Myroslav Opyr |
|
From: Myroslav O. <My...@or...> - 2002-03-21 01:38:24
|
Morten W. Petersen wrote:
>>What interest me now is complex screens with different functional
>>(informational) parts on the same page.
>>
>Cool, that's one of the areas I'd like to look into as well. :)
>
A couple of statements and then questions.
Compound pages should display not only one specific flavor of object but
many. Potential candidates for Building Blocks are:
* object properties (html_edit_properties)
* subobjects (object_manager_index, with different meta_type)
* search
* feedback
* history
* help
* custom blocks
Questions:
* What are other suggested Building Blocks?
* What are types of layout?
* Should layout be smart to guess what and where to display?
* On what level layout should be defined? in Python code? DTML in
FS? DTML in ZODB? Layout defiition in ZODB?
Just thoughts:
One of the layouts I am "drawing" is grid layout. Table with two columns
and multiple rows. Each cell contain building block. Most reasonable are
Properties and subobjects (as many as get_meta_types() returns). Other
obvious layouts are similar layouts without fixed vertical or horizontal
margins. "without vertical margin" is table with one cell per row but
each cell contains two buildings blocks (they'll decide how to divide
space). "without vertical" is similar but in other direction.
Regards,
Myroslav
|
|
From: Myroslav O. <My...@or...> - 2002-03-20 22:54:03
|
Morten W. Petersen wrote: >On Wed, Mar 20, 2002 at 12:25:18AM +0200, Myroslav Opyr wrote: > >>Morten W. Petersen wrote: >> >>>Why are friendly URLs essential? >>> >>It is human-factor. >> >> 1. When I look at friendly url I can guess what is going on even when >> I know nothing about the system. >> 2. Statistical analysis (webalizer output) of application (for >> usability improvements) is much easier if I see readable URLs. >> 3. There is no need to use extra tools for crash analysis except logs >> with friendly URLs. >> 4. Other options I have no imagination now, but they exists. >> >OK, I'd like to see a 'spike' (example code) for this, to see how it >could be implemented. > First I'd like to remove Zope ID dependency. To achieve this I'd add immutable warp_id (or wid to shorten) to each object within WF application (within scope of id_generator). Upon creation (or clonong) each WF object should get an Unique warp_id and register itself in Catalog. Later all refernces within application should be warp_id based. Catalog would be the engine which resolve all such references. The described above change is not spike scope one. It is major system modification. catalog.py and all get_catalog() related code should be modified at least... In the beginning I'd left ID generation as is but later it would change to allow arbitrary id changes and object location within object hierarchy (as soon as the object updates Catalog after such moves). Myroslav |
|
From: Myroslav O. <My...@or...> - 2002-03-20 22:34:40
|
Morten W. Petersen wrote: > On Wed, Mar 20, 2002 at 12:39:23AM +0200, Myroslav Opyr wrote: > >> What if there will be some way to convert wf-based aplication to UML >> and back? I was analysing my actions. First I've drawn a UML schema >> then I silly translated it into Python code. Adding only functions >> limiting datasets (discussed in "Q: Dynamic queries" thread) and >> functions generating vatious titles of objects... >> > There's already a program that produces python classes from UML, called > dia2code [1]. Haven't tested it so I'm not sure how it works. > I'll take a look. > That could be modified to output 'WF frienly code' (with properties), > and some methods in WF (like render_class_hierarchy) could output UML. > What if the whole system moves to use of interfaces? It look like that framework would be easily and natively ported to Zope3 (hopefully together with applications based on it). > This all brings up an interesting question, and that is will WF support > 'workflow'-like behaviour (UML Activity Diagrams)? I've been > contemplating creating a basic workflow system (wizard-like interfaces) > for WF, do you have any thoughts on that? > I am not on that stage of problem domain understanding. I have another general idea about whole WF (it is subject of another thread). >> P.S. I've found render_class_hierarchy method what other interesting >> things are there? >> > There's nothing similar in WF that I know of. :) > It's a pity. It is fun to find such "utilities". |
|
From: Myroslav O. <My...@or...> - 2002-03-20 22:34:40
|
Morten W. Petersen wrote: > On Tue, Mar 19, 2002 at 11:06:03PM +0200, Myroslav Opyr wrote: > >> Today I got strange (for me at least) compatibility problems. I tried >> to move warp-framework based application from one Zope instance to >> another one. Additionally it was platform change, I moved from Win32 >> to FreeBSD4. I used zexp for "migration". I partly solved the problem >> with * * thread_safe_counter (recreated it's file in var dir) and >> having updated Catalog. But I got stuck with problems appearing when >> instance of any object is created. It just refuses to render the page >> (object_manager_index containing the object). Object is creatred and >> deleting it from ZMI solves the problem. > > What if you create a new instance on FreeBSD4 and add data to it, does > that work? Yes it works. I've created another instance. None of the symptoms above... I'll continue experiments to find what is causing the problem. Myroslav |
|
From: Myroslav O. <My...@or...> - 2002-03-20 22:14:27
|
Morten W. Petersen wrote: >On Tue, Mar 19, 2002 at 11:54:28PM +0200, Myroslav Opyr wrote: > >>Morten W. Petersen wrote: >> >>>I don't see any good reason to log errors in two places... >>> >>STUPID_LOG_FILE is huge and on production servers it is not enabled. >>Thus sometimes it is necessary to know the source. It is possible to ask >>customer to "View Source", Copy, Paste and mail it to me. And having >>several errors on one page sometimes it is difficult to distiguish >>single one necessary out of log. >> >What if the error is mailed automatically? With path to object, >type of object, error and traceback? > For production system it is reasonable. But I'd become mad if that feature was active with my appication in it's current state ;) Opinion: use standard MailHost object to send/receive mail instead of direct smtplib. It would be another hook point. Objections? Thinking further: registration, notification, logging and paging. Why not create some system (equivalent of syslog) and implement several scenarios: such as logging into mutiple files, notification via e-mail (SMS, pager), registration into database (in a journal of user activities, in on row with registration of user pageviews currently present in WF). Does that sound reasonable? Myroslav |
|
From: Morten W. P. <mo...@th...> - 2002-03-20 20:55:20
|
On Wed, Mar 20, 2002 at 12:39:23AM +0200, Myroslav Opyr wrote: > What if there will be some way to convert wf-based aplication to UML and > back? I was analysing my actions. First I've drawn a UML schema then I > silly translated it into Python code. Adding only functions limiting > datasets (discussed in "Q: Dynamic queries" thread) and functions > generating vatious titles of objects... There's already a program that produces python classes from UML, called dia2code [1]. Haven't tested it so I'm not sure how it works. That could be modified to output 'WF frienly code' (with properties), and some methods in WF (like render_class_hierarchy) could output UML. This all brings up an interesting question, and that is will WF support 'workflow'-like behaviour (UML Activity Diagrams)? I've been contemplating creating a basic workflow system (wizard-like interfaces) for WF, do you have any thoughts on that? > P.S. I've found render_class_hierarchy method what other interesting > things are there? There's nothing similar in WF that I know of. :) [1] http://sourceforge.net/projects/dia2code/ -Morten |
|
From: Morten W. P. <mo...@th...> - 2002-03-20 20:22:52
|
On Tue, Mar 19, 2002 at 11:54:28PM +0200, Myroslav Opyr wrote: > > Morten W. Petersen wrote: > > >I don't see any good reason to log errors in two places... > > > STUPID_LOG_FILE is huge and on production servers it is not enabled. > Thus sometimes it is necessary to know the source. It is possible to ask > customer to "View Source", Copy, Paste and mail it to me. And having > several errors on one page sometimes it is difficult to distiguish > single one necessary out of log. What if the error is mailed automatically? With path to object, type of object, error and traceback? -Morten |
|
From: Morten W. P. <mo...@th...> - 2002-03-20 20:21:43
|
On Wed, Mar 20, 2002 at 12:25:18AM +0200, Myroslav Opyr wrote: > Morten W. Petersen wrote: > > >Why are friendly URLs essential? > > > It is human-factor. > > 1. When I look at friendly url I can guess what is going on even when > I know nothing about the system. > 2. Statistical analysis (webalizer output) of application (for > usability improvements) is much easier if I see readable URLs. > 3. There is no need to use extra tools for crash analysis except logs > with friendly URLs. > 4. Other options I have no imagination now, but they exists. OK, I'd like to see a 'spike' (example code) for this, to see how it could be implemented. -Morten |
1 message has been excluded from this view by a project administrator.
