thats pretty amazing that PyQt can handle such a big dataset at once. Putting JavaScript aside just trying to load html tables with a large amount of rows can become very unwieldy. Good luck with the app.
cheers, Marwan On Fri, Jun 17, 2011 at 5:55 PM, Knack <knack...@googlemail.com> wrote: > The requirements are: > > 1) Handling large (>100000 lines, ~ 10 columns) amounts of structured > data in treeviews. For certain use cases comparing two of those trees. > 2) Creating different structured views on the data. Like using > different categories of attributes as structure criteria. Those views > are not persistent. > 3) Performing filter and configuration operations on the data. Those > can be quite expensive, for example variant configuration which comes > down to evaluating boolean expressions for each item. > 4) Maybe (if there's time and I find an easy to use lib, maybe > Panda3D) adding a simple 3D viewer. > > The decision has been made without really evaluation JavaScript or > Flash. I must admit that I don't know wether the state of the art JS > libraries can handle that large trees. But I know that PyQt can handle > that large data amazingly well. Further, I have some experience with > these kind of systems: Most of them are quite "server-centric" (as > common web apps are). Those systems have all performance issues when > the user needs another view or configuration of the data. My approach > is to eagerly load all the data to the client and do all calculations > for the views there. I must add that it's an in-house app and all > users have proper PCs. > > Lastly, Personal experience and know-how: PyQt is fun to work with > (after some learning curve, of course). It has style-sheets for making > the client sexy at the end. On the other hand, I'm totally impressed > what can be done with JS, but I'm not familiar with it. > > Cheers, > > Jan > > > > On 17 Jun., 06:26, Marwan Al-Sabbagh <marwan.alsabb...@gmail.com> > wrote: > > Out of curiosity is there a specific need for running a rich client > instead > > of just accessing the system app through a web browser. > > > > > > > > > > > > > > > > On Fri, Jun 10, 2011 at 4:16 AM, Ovnicraft <ovnicr...@gmail.com> wrote: > > > soaplib <https://github.com/soaplib/> > > > > > On Thu, Jun 9, 2011 at 1:10 PM, Knack <knack...@googlemail.com> wrote: > > > > >> Hi guys, > > > > >> I've got a prototype DB + Rich Client app and need to make a proper > > >> production app out of it. The client is written in PyQt and needs to > > >> fullfill quite some functional and performance requirements (I would > > >> really like to keep it). The client connects directly to the DB. > > > > >> After some thoughts about security I came to the conclusion that using > > >> django is the way to go. Now I seek some coarse advice / feedback > > >> wether I'm heading in the right directions. I've got some miles to go > > >> and looking forward to learning Django, but want to avoid newbie > > >> pitfalls. So if something of the following sounds like bad ideas to > > >> you, please let me know ;-). > > > > >> Current plans are: > > > > >> 1) Using an Oracle DB > > >> 2) LDAP authentification > > >> 3) Role based authorisation. Here I'm a bit unsure about the approach. > > >> I would implement it by myself within the django world (don't know yet > > >> where exactly). > > >> 4) Transfering the data using a web service via https. Here are my > > >> biggest uncertainties. What kind of web service to use? Rest, xml-rpc > > > > > I recommend you REsT, IMHO XML-RPC is not fast, SOAP is a good option > but > > > we get just few libraries under development soaplib[1] is really > promise. > > > > > or soap? How to package the acual data? Embedded in the html or xml? > > > > > Just be light. > > > > > Regards, > > > > > As binary payload? I would need only basic stuff like log in, log out, > > >> read and write whole database rows, which are basically the business > > >> objects. Most of the logic is done on the rich client. I would > > >> transfer large amounts (100000s of business objects) of data at > > >> startup time (very eager loading to the client). But also lazy loading > > >> of course. > > >> 5) One important concept is that I change only a small amount of the > > >> data objects. Most of the changes lead to new versions. Therefore I > > >> hope to be able to use caching extensivly. The data objects which are > > >> immutable could be permanently cached in memory. How about > > >> authorisation of these cached objects? Are there ways django takes > > >> care of this? > > > > >> A long first post ;-). Any comments are appreciated! > > > > >> Cheers, > > > > >> Jan > > > > >> -- > > >> You received this message because you are subscribed to the Google > Groups > > >> "Django users" group. > > >> To post to this group, send email to django-users@googlegroups.com. > > >> To unsubscribe from this group, send email to > > >> django-users+unsubscr...@googlegroups.com. > > >> For more options, visit this group at > > >>http://groups.google.com/group/django-users?hl=en. > > > > > -- > > > Cristian Salamea > > > @ovnicraft > > > > > -- > > > You received this message because you are subscribed to the Google > Groups > > > "Django users" group. > > > To post to this group, send email to django-users@googlegroups.com. > > > To unsubscribe from this group, send email to > > > django-users+unsubscr...@googlegroups.com. > > > For more options, visit this group at > > >http://groups.google.com/group/django-users?hl=en. > > -- > You received this message because you are subscribed to the Google Groups > "Django users" group. > To post to this group, send email to django-users@googlegroups.com. > To unsubscribe from this group, send email to > django-users+unsubscr...@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/django-users?hl=en. > > -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-users@googlegroups.com. To unsubscribe from this group, send email to django-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/django-users?hl=en.
