This presentation talks about Django, but there are things that may be useful for the evolution of Flask: http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it IMHO Flask do it better than Django, and it is written also there. The suggestion about models is a good thing to think about (I will do). Actually I fix the configuration of a Flask app moving some function calls from Flask.__init__(). What do you think? Davide Muzzarelli www.dav-muz.net
I don't think Models should be part of a framework because nobody really get's them right (in my opinion). The approach frameworks take to models is usually too low level so these models can be used only as part of an ORM and form stuff is somehow built around that if at all. Usually you basically end up repeating yourself for forms and REST APIs. A Model should be high level enough to allow any kind of datastorage backend, a relational database, an object oriented database, a document oriented database, a key/value store or a file which could be used for configuration. Especially given simple files as backends for configurations it should obviously be possible to choose a backend on a per model basis. Also models should be aware of public, private and internal(ForeignKey's in SQLAlchemy as opposed to relations which would probably be public) fields, which can be used to create forms and serialize the data structure into formats such as JSON, YAML, XML or whatever else might be useful for Queues used to implement asynchronous tasks or simply a REST API. Another important feature would be validation directly in the model as opposed to validation in forms, which minimizes the risk of failure by putting the validation as low-level as possible.
In data giovedì 09 settembre 2010 20:18:42, Daniel Neuhäuser ha scritto: > A Model should be high level enough to allow any kind of datastorage > backend, a relational database, an object oriented database, a document > oriented database, a key/value store or a file which could be used for > configuration. IMHO make a model compatible with different SQL backends or libraries is a possible way. I'ts a good thing if don't waste performances and developer's time. Always need a solution. Instead, make a model compatible with different types of database (SQL, noSQL, dict etc...) is an impossible way because each type have different working methods. > Another important feature would be validation directly in the model > as opposed to validation in forms, which minimizes the risk of failure > by putting the validation as low-level as possible. I think not because the forms are for the user and not for the models. Between users and models there are views that convert what the user want and how to do the things. A single form could do logic and create datas using different models at the same time. Often a view can create a complex system with a simple form with few field. Coupling forms to models restrict you to use only that form library and that database utility, also if you don't want. Coupling form restrict you to use a model in only one predefined way. This is why Django's Admin was rewritten: because Django couple forms and models (more in the past than now). I understand that coupling forms with models make some things faster to write, but in close time this is a waste of time when you want to replace something. Please, tell me if I don't understand well your suggestion. Davide Muzzarelli www.dav-muz.net
I am not so much of a fan of built in models. Built in model classes imply the use of an ORM which then leads you to possibly a SQL-only ORM, which then could become tied into every other aspect of the framework. This is what killed Django for me, over time. The more that we started using other database engines, and things, Django's ORM became such a crutch that we decided to look for alternatives - we found Flask. On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli <d.muzzarelli@dav-muz.net > wrote: > This presentation talks about Django, but there are things that may be > useful > for the evolution of Flask: > http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it > > IMHO Flask do it better than Django, and it is written also there. > > The suggestion about models is a good thing to think about (I will do). > Actually I fix the configuration of a Flask app moving some function calls > from > Flask.__init__(). > > What do you think? > > Davide Muzzarelli > www.dav-muz.net >
I think Django has too much CMS heritage in its design; the predefined auth models, the admin app, and so on. That's not to diminish its usefulness for certain projects, and it's certainly improved over the last few years. But I dislike the monoculture of Python web development (almost every Python web dev job you see nowadays is using Django) so it's good to be able to offer alternatives - Flask has a more minimalist, "use what you need" philosophy which is more suited to more original projects. On 9 September 2010 16:44, Ryan Cole <betawarz@gmail.com> wrote: > I am not so much of a fan of built in models. Built in model classes imply > the use of an ORM which then leads you to possibly a SQL-only ORM, which > then could become tied into every other aspect of the framework. This is > what killed Django for me, over time. The more that we started using other > database engines, and things, Django's ORM became such a crutch that we > decided to look for alternatives - we found Flask. > On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli > <d.muzzarelli@dav-muz.net> wrote: >> >> This presentation talks about Django, but there are things that may be >> useful >> for the evolution of Flask: >> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >> >> IMHO Flask do it better than Django, and it is written also there. >> >> The suggestion about models is a good thing to think about (I will do). >> Actually I fix the configuration of a Flask app moving some function calls >> from >> Flask.__init__(). >> >> What do you think? >> >> Davide Muzzarelli >> www.dav-muz.net > >
> But I dislike the monoculture of Python web development
That is an incredible statement. You meant to type that?
Python web dev is a vast, loosely connected cloud of frameworks,
difficult to evaluate, since Pypi is lacking in some areas, although
they did have 'kwalitee.' djangopackages.com and readthedocs.org
will help greatly with evalutation of the numerous python packages.
I'd say that Django has excellent and extensive documentation. They
also market themselves better, a la github. This is why IMO Django is
a common job requirement.
The big takeaway for me was the benchmark idea. Love it. Going to
implement it as many places as I can.
On 10 September 2010 09:10, kevin beckford <chiggsy@lazyweb.ca> wrote: >> But I dislike the monoculture of Python web development > > That is an incredible statement. You meant to type that? > Yes. I'm referring to the fact that almost all Python web dev jobs right now are Django jobs, and it's become the defacto Python web framework for new projects. I am not referring to the number of existing Python web frameworks. > > I'd say that Django has excellent and extensive documentation. They > also market themselves better, a la github. This is why IMO Django is > a common job requirement. That may well be the case, if so it's worth challenging with better documentation and high profile projects.
At least it's got some options, like Flask, Pylons, TurboGears, etc. In Ruby, it's Rails or nothing. On Sep 9, 2010, at 12:04 PM, Dan Jacob wrote: > I think Django has too much CMS heritage in its design; the predefined > auth models, the admin app, and so on. That's not to diminish its > usefulness for certain projects, and it's certainly improved over the > last few years. But I dislike the monoculture of Python web > development (almost every Python web dev job you see nowadays is using > Django) so it's good to be able to offer alternatives - Flask has a > more minimalist, "use what you need" philosophy which is more suited > to more original projects. > > On 9 September 2010 16:44, Ryan Cole <betawarz@gmail.com> wrote: >> I am not so much of a fan of built in models. Built in model classes imply >> the use of an ORM which then leads you to possibly a SQL-only ORM, which >> then could become tied into every other aspect of the framework. This is >> what killed Django for me, over time. The more that we started using other >> database engines, and things, Django's ORM became such a crutch that we >> decided to look for alternatives - we found Flask. >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli >> <d.muzzarelli@dav-muz.net> wrote: >>> >>> This presentation talks about Django, but there are things that may be >>> useful >>> for the evolution of Flask: >>> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >>> >>> IMHO Flask do it better than Django, and it is written also there. >>> >>> The suggestion about models is a good thing to think about (I will do). >>> Actually I fix the configuration of a Flask app moving some function calls >>> from >>> Flask.__init__(). >>> >>> What do you think? >>> >>> Davide Muzzarelli >>> www.dav-muz.net >> >>
At least it's got some options, like Flask, Pylons, TurboGears, etc. In Ruby, it's Rails or nothing. On Sep 9, 2010, at 12:04 PM, Dan Jacob wrote: > I think Django has too much CMS heritage in its design; the predefined > auth models, the admin app, and so on. That's not to diminish its > usefulness for certain projects, and it's certainly improved over the > last few years. But I dislike the monoculture of Python web > development (almost every Python web dev job you see nowadays is using > Django) so it's good to be able to offer alternatives - Flask has a > more minimalist, "use what you need" philosophy which is more suited > to more original projects. > > On 9 September 2010 16:44, Ryan Cole <betawarz@gmail.com> wrote: >> I am not so much of a fan of built in models. Built in model classes imply >> the use of an ORM which then leads you to possibly a SQL-only ORM, which >> then could become tied into every other aspect of the framework. This is >> what killed Django for me, over time. The more that we started using other >> database engines, and things, Django's ORM became such a crutch that we >> decided to look for alternatives - we found Flask. >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli >> <d.muzzarelli@dav-muz.net> wrote: >>> >>> This presentation talks about Django, but there are things that may be >>> useful >>> for the evolution of Flask: >>> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >>> >>> IMHO Flask do it better than Django, and it is written also there. >>> >>> The suggestion about models is a good thing to think about (I will do). >>> Actually I fix the configuration of a Flask app moving some function calls >>> from >>> Flask.__init__(). >>> >>> What do you think? >>> >>> Davide Muzzarelli >>> www.dav-muz.net >> >>
For Ruby there is also Sintara. http://www.sinatrarb.com/ I would really appreciate not using any kind of built in models (SQL-only ORM). I'm mostly using Flask on App Engine and it fits perfect, cause Flask is covering exactly the parts a framework should handle (on my opinion) On Sep 10, 2010, at 12:30 AM, Nolan Brubaker wrote: > At least it's got some options, like Flask, Pylons, TurboGears, etc. > > In Ruby, it's Rails or nothing. > > > On Sep 9, 2010, at 12:04 PM, Dan Jacob wrote: > >> I think Django has too much CMS heritage in its design; the predefined >> auth models, the admin app, and so on. That's not to diminish its >> usefulness for certain projects, and it's certainly improved over the >> last few years. But I dislike the monoculture of Python web >> development (almost every Python web dev job you see nowadays is using >> Django) so it's good to be able to offer alternatives - Flask has a >> more minimalist, "use what you need" philosophy which is more suited >> to more original projects. >> >> On 9 September 2010 16:44, Ryan Cole <betawarz@gmail.com> wrote: >>> I am not so much of a fan of built in models. Built in model classes imply >>> the use of an ORM which then leads you to possibly a SQL-only ORM, which >>> then could become tied into every other aspect of the framework. This is >>> what killed Django for me, over time. The more that we started using other >>> database engines, and things, Django's ORM became such a crutch that we >>> decided to look for alternatives - we found Flask. >>> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli >>> <d.muzzarelli@dav-muz.net> wrote: >>>> >>>> This presentation talks about Django, but there are things that may be >>>> useful >>>> for the evolution of Flask: >>>> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >>>> >>>> IMHO Flask do it better than Django, and it is written also there. >>>> >>>> The suggestion about models is a good thing to think about (I will do). >>>> Actually I fix the configuration of a Flask app moving some function calls >>>> from >>>> Flask.__init__(). >>>> >>>> What do you think? >>>> >>>> Davide Muzzarelli >>>> www.dav-muz.net >>> >>> >
On 9 September 2010 23:30, Nolan Brubaker <nolan.brubaker@gmail.com> wrote: > At least it's got some options, like Flask, Pylons, TurboGears, etc. > > In Ruby, it's Rails or nothing. > > What about Sinatra, Camping and Merb ? > On Sep 9, 2010, at 12:04 PM, Dan Jacob wrote: > >> I think Django has too much CMS heritage in its design; the predefined >> auth models, the admin app, and so on. That's not to diminish its >> usefulness for certain projects, and it's certainly improved over the >> last few years. But I dislike the monoculture of Python web >> development (almost every Python web dev job you see nowadays is using >> Django) so it's good to be able to offer alternatives - Flask has a >> more minimalist, "use what you need" philosophy which is more suited >> to more original projects. >> >> On 9 September 2010 16:44, Ryan Cole <betawarz@gmail.com> wrote: >>> I am not so much of a fan of built in models. Built in model classes imply >>> the use of an ORM which then leads you to possibly a SQL-only ORM, which >>> then could become tied into every other aspect of the framework. This is >>> what killed Django for me, over time. The more that we started using other >>> database engines, and things, Django's ORM became such a crutch that we >>> decided to look for alternatives - we found Flask. >>> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli >>> <d.muzzarelli@dav-muz.net> wrote: >>>> >>>> This presentation talks about Django, but there are things that may be >>>> useful >>>> for the evolution of Flask: >>>> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >>>> >>>> IMHO Flask do it better than Django, and it is written also there. >>>> >>>> The suggestion about models is a good thing to think about (I will do). >>>> Actually I fix the configuration of a Flask app moving some function calls >>>> from >>>> Flask.__init__(). >>>> >>>> What do you think? >>>> >>>> Davide Muzzarelli >>>> www.dav-muz.net >>> >>> > >
Fair enough on Sinatra and Camping. Is Merb still available as an independent project, though? I thought Merb was Rails 3 now. In any case, I've been able to get Flask used on our projects at work - we're really impressed with the project, and for what we're doing right now (RESTful P2P library/application), it fits way better for that than the bigger frameworks. On Sep 9, 2010, at 6:54 PM, Dan Jacob wrote: > On 9 September 2010 23:30, Nolan Brubaker <nolan.brubaker@gmail.com> wrote: >> At least it's got some options, like Flask, Pylons, TurboGears, etc. >> >> In Ruby, it's Rails or nothing. >> >> > > > What about Sinatra, Camping and Merb ? >
I completely agree with Ryan. However, I can't see the harm in someone making an unofficial distribution of flask with a baked in orm/model system. It could possibly increase adoption. Question is, do we even want those people using Flask? Likely not. Someone who wants a Django-esche framework with more flexibility should likely start with Pylons. If someone wants to build an application completely tailored to their needs, then they should use Flask. Kenneth Reitz http://kennethreitz.com/contact-me On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> wrote: > I am not so much of a fan of built in models. Built in model classes imply > the use of an ORM which then leads you to possibly a SQL-only ORM, which > then could become tied into every other aspect of the framework. This is > what killed Django for me, over time. The more that we started using other > database engines, and things, Django's ORM became such a crutch that we > decided to look for alternatives - we found Flask. > > On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli < > d.muzzarelli@dav-muz.net> wrote: > >> This presentation talks about Django, but there are things that may be >> useful >> for the evolution of Flask: >> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >> >> IMHO Flask do it better than Django, and it is written also there. >> >> The suggestion about models is a good thing to think about (I will do). >> Actually I fix the configuration of a Flask app moving some function calls >> from >> Flask.__init__(). >> >> What do you think? >> >> Davide Muzzarelli >> www.dav-muz.net >> > >
Just my two cents, I started using Flask because it was extremely minimal and left all the choices up to me. IMO, that is the spirit of Flask. --Dan On Sep 9, 2010, at 9:00 AM, Kenneth Reitz wrote: > I completely agree with Ryan. > > However, I can't see the harm in someone making an unofficial distribution of flask with a baked in orm/model system. It could possibly increase adoption. > > Question is, do we even want those people using Flask? Likely not. Someone who wants a Django-esche framework with more flexibility should likely start with Pylons. > > If someone wants to build an application completely tailored to their needs, then they should use Flask. > > > Kenneth Reitz > http://kennethreitz.com/contact-me > > > On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> wrote: > I am not so much of a fan of built in models. Built in model classes imply the use of an ORM which then leads you to possibly a SQL-only ORM, which then could become tied into every other aspect of the framework. This is what killed Django for me, over time. The more that we started using other database engines, and things, Django's ORM became such a crutch that we decided to look for alternatives - we found Flask. > > On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli <d.muzzarelli@dav-muz.net> wrote: > This presentation talks about Django, but there are things that may be useful > for the evolution of Flask: > http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it > > IMHO Flask do it better than Django, and it is written also there. > > The suggestion about models is a good thing to think about (I will do). > Actually I fix the configuration of a Flask app moving some function calls from > Flask.__init__(). > > What do you think? > > Davide Muzzarelli > www.dav-muz.net > >
Exactly.....when I see the word django I close the page and move on. For me, the "mocroness" of flask is why I love it ! Sent from my iPhone On 9 Sep 2010, at 17:11, Dan Colish <dcolish@gmail.com> wrote: > Just my two cents, I started using Flask because it was extremely > minimal and left all the choices up to me. IMO, that is the spirit > of Flask. > > --Dan > > On Sep 9, 2010, at 9:00 AM, Kenneth Reitz wrote: > >> I completely agree with Ryan. >> >> However, I can't see the harm in someone making an unofficial >> distribution of flask with a baked in orm/model system. It could >> possibly increase adoption. >> >> Question is, do we even want those people using Flask? Likely not. >> Someone who wants a Django-esche framework with more flexibility >> should likely start with Pylons. >> >> If someone wants to build an application completely tailored to >> their needs, then they should use Flask. >> >> >> Kenneth Reitz >> http://kennethreitz.com/contact-me >> >> >> On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> >> wrote: >> I am not so much of a fan of built in models. Built in model >> classes imply the use of an ORM which then leads you to possibly a >> SQL-only ORM, which then could become tied into every other aspect >> of the framework. This is what killed Django for me, over time. The >> more that we started using other database engines, and things, >> Django's ORM became such a crutch that we decided to look for >> alternatives - we found Flask. >> >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli <d.muzzarelli@dav-muz.net >> > wrote: >> This presentation talks about Django, but there are things that may >> be useful >> for the evolution of Flask: >> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >> >> IMHO Flask do it better than Django, and it is written also there. >> >> The suggestion about models is a good thing to think about (I will >> do). >> Actually I fix the configuration of a Flask app moving some >> function calls from >> Flask.__init__(). >> >> What do you think? >> >> Davide Muzzarelli >> www.dav-muz.net >> >> >
+1 On Sep 9, 2010 12:27 PM, "Dan Colish" <dcolish@gmail.com> wrote: > Just my two cents, I started using Flask because it was extremely minimal and left all the choices up to me. IMO, that is the spirit of Flask. > > --Dan > > On Sep 9, 2010, at 9:00 AM, Kenneth Reitz wrote: > >> I completely agree with Ryan. >> >> However, I can't see the harm in someone making an unofficial distribution of flask with a baked in orm/model system. It could possibly increase adoption. >> >> Question is, do we even want those people using Flask? Likely not. Someone who wants a Django-esche framework with more flexibility should likely start with Pylons. >> >> If someone wants to build an application completely tailored to their needs, then they should use Flask. >> >> >> Kenneth Reitz >> http://kennethreitz.com/contact-me >> >> >> On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> wrote: >> I am not so much of a fan of built in models. Built in model classes imply the use of an ORM which then leads you to possibly a SQL-only ORM, which then could become tied into every other aspect of the framework. This is what killed Django for me, over time. The more that we started using other database engines, and things, Django's ORM became such a crutch that we decided to look for alternatives - we found Flask. >> >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli < d.muzzarelli@dav-muz.net> wrote: >> This presentation talks about Django, but there are things that may be useful >> for the evolution of Flask: >> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >> >> IMHO Flask do it better than Django, and it is written also there. >> >> The suggestion about models is a good thing to think about (I will do). >> Actually I fix the configuration of a Flask app moving some function calls from >> Flask.__init__(). >> >> What do you think? >> >> Davide Muzzarelli >> www.dav-muz.net >> >> >
I don't do a lot of 'real' web stuff, but I like how flask is now. I'm re-learning how to do web dev these days and tried Django first. For most of the things I'm doing I don't need a lot of the 'magic'. Most of what I do is turning various SA tasks into something web-based and simplify things. I don't need an ORM for something that needs 2 or 3 queries, or back end administration. I have nothing against an extension adding this functionality, but I don't want to *have* to use something. Tim On Thu, Sep 9, 2010 at 12:32 PM, Kenneth Reitz <me@kennethreitz.com> wrote: > +1 > > On Sep 9, 2010 12:27 PM, "Dan Colish" <dcolish@gmail.com> wrote: > > Just my two cents, I started using Flask because it was extremely minimal > and left all the choices up to me. IMO, that is the spirit of Flask. > > > > --Dan > > > > On Sep 9, 2010, at 9:00 AM, Kenneth Reitz wrote: > > > >> I completely agree with Ryan. > >> > >> However, I can't see the harm in someone making an unofficial > distribution of flask with a baked in orm/model system. It could possibly > increase adoption. > >> > >> Question is, do we even want those people using Flask? Likely not. > Someone who wants a Django-esche framework with more flexibility should > likely start with Pylons. > >> > >> If someone wants to build an application completely tailored to their > needs, then they should use Flask. > >> > >> > >> Kenneth Reitz > >> http://kennethreitz.com/contact-me > >> > >> > >> On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> wrote: > >> I am not so much of a fan of built in models. Built in model classes > imply the use of an ORM which then leads you to possibly a SQL-only ORM, > which then could become tied into every other aspect of the framework. This > is what killed Django for me, over time. The more that we started using > other database engines, and things, Django's ORM became such a crutch that > we decided to look for alternatives - we found Flask. > >> > >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli < > d.muzzarelli@dav-muz.net> wrote: > >> This presentation talks about Django, but there are things that may be > useful > >> for the evolution of Flask: > >> > http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it > >> > >> IMHO Flask do it better than Django, and it is written also there. > >> > >> The suggestion about models is a good thing to think about (I will do). > >> Actually I fix the configuration of a Flask app moving some function > calls from > >> Flask.__init__(). > >> > >> What do you think? > >> > >> Davide Muzzarelli > >> www.dav-muz.net > >> > >> > > >
> > On Thu, Sep 9, 2010 at 2:36 PM, Tim Conrad <tim@timconrad.org> wrote: > > [...] > I have nothing against an extension adding this functionality, but I don't > want to *have* to use something. > +1 :) Cheers, Francisco Souza Software developer at Giran and also full time Open source evangelist at full time English: http://www.franciscosouza.net Portuguese: http://www.franciscosouza.com.br Twitter: @franciscosouza +55 27 3026 0264 On Thu, Sep 9, 2010 at 2:36 PM, Tim Conrad <tim@timconrad.org> wrote: > I don't do a lot of 'real' web stuff, but I like how flask is now. I'm > re-learning how to do web dev these days and tried Django first. For most of > the things I'm doing I don't need a lot of the 'magic'. Most of what I do is > turning various SA tasks into something web-based and simplify things. I > don't need an ORM for something that needs 2 or 3 queries, or back end > administration. > > I have nothing against an extension adding this functionality, but I don't > want to *have* to use something. > > Tim > > > On Thu, Sep 9, 2010 at 12:32 PM, Kenneth Reitz <me@kennethreitz.com>wrote: > >> +1 >> >> On Sep 9, 2010 12:27 PM, "Dan Colish" <dcolish@gmail.com> wrote: >> > Just my two cents, I started using Flask because it was extremely >> minimal and left all the choices up to me. IMO, that is the spirit of Flask. >> > >> > --Dan >> > >> > On Sep 9, 2010, at 9:00 AM, Kenneth Reitz wrote: >> > >> >> I completely agree with Ryan. >> >> >> >> However, I can't see the harm in someone making an unofficial >> distribution of flask with a baked in orm/model system. It could possibly >> increase adoption. >> >> >> >> Question is, do we even want those people using Flask? Likely not. >> Someone who wants a Django-esche framework with more flexibility should >> likely start with Pylons. >> >> >> >> If someone wants to build an application completely tailored to their >> needs, then they should use Flask. >> >> >> >> >> >> Kenneth Reitz >> >> http://kennethreitz.com/contact-me >> >> >> >> >> >> On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> wrote: >> >> I am not so much of a fan of built in models. Built in model classes >> imply the use of an ORM which then leads you to possibly a SQL-only ORM, >> which then could become tied into every other aspect of the framework. This >> is what killed Django for me, over time. The more that we started using >> other database engines, and things, Django's ORM became such a crutch that >> we decided to look for alternatives - we found Flask. >> >> >> >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli < >> d.muzzarelli@dav-muz.net> wrote: >> >> This presentation talks about Django, but there are things that may be >> useful >> >> for the evolution of Flask: >> >> >> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >> >> >> >> IMHO Flask do it better than Django, and it is written also there. >> >> >> >> The suggestion about models is a good thing to think about (I will do). >> >> Actually I fix the configuration of a Flask app moving some function >> calls from >> >> Flask.__init__(). >> >> >> >> What do you think? >> >> >> >> Davide Muzzarelli >> >> www.dav-muz.net >> >> >> >> >> > >> > >
Flask-SQLAlchemy, Flask-MongoDB, Flask-Couch Done. I think Flask for the most part is complete, everything else should be left up to extensions. I'd almost like to go so far as to say that jinja2 shouldn't be configured by default... but a Flask-Jinja2 extension created instead.... -- Thadeus On Thu, Sep 9, 2010 at 1:09 PM, Francisco Souza <francisco@franciscosouza.net> wrote: >> On Thu, Sep 9, 2010 at 2:36 PM, Tim Conrad <tim@timconrad.org> wrote: >> >> [...] >> I have nothing against an extension adding this functionality, but I don't >> want to *have* to use something. > > +1 :) > > Cheers, > Francisco Souza > Software developer at Giran and also full time > Open source evangelist at full time > > English: http://www.franciscosouza.net > Portuguese: http://www.franciscosouza.com.br > Twitter: @franciscosouza > +55 27 3026 0264 > > > On Thu, Sep 9, 2010 at 2:36 PM, Tim Conrad <tim@timconrad.org> wrote: >> >> I don't do a lot of 'real' web stuff, but I like how flask is now. I'm >> re-learning how to do web dev these days and tried Django first. For most of >> the things I'm doing I don't need a lot of the 'magic'. Most of what I do is >> turning various SA tasks into something web-based and simplify things. I >> don't need an ORM for something that needs 2 or 3 queries, or back end >> administration. >> I have nothing against an extension adding this functionality, but I don't >> want to *have* to use something. >> Tim >> >> On Thu, Sep 9, 2010 at 12:32 PM, Kenneth Reitz <me@kennethreitz.com> >> wrote: >>> >>> +1 >>> >>> On Sep 9, 2010 12:27 PM, "Dan Colish" <dcolish@gmail.com> wrote: >>> > Just my two cents, I started using Flask because it was extremely >>> > minimal and left all the choices up to me. IMO, that is the spirit of Flask. >>> > >>> > --Dan >>> > >>> > On Sep 9, 2010, at 9:00 AM, Kenneth Reitz wrote: >>> > >>> >> I completely agree with Ryan. >>> >> >>> >> However, I can't see the harm in someone making an unofficial >>> >> distribution of flask with a baked in orm/model system. It could possibly >>> >> increase adoption. >>> >> >>> >> Question is, do we even want those people using Flask? Likely not. >>> >> Someone who wants a Django-esche framework with more flexibility should >>> >> likely start with Pylons. >>> >> >>> >> If someone wants to build an application completely tailored to their >>> >> needs, then they should use Flask. >>> >> >>> >> >>> >> Kenneth Reitz >>> >> http://kennethreitz.com/contact-me >>> >> >>> >> >>> >> On Thu, Sep 9, 2010 at 11:44 AM, Ryan Cole <betawarz@gmail.com> wrote: >>> >> I am not so much of a fan of built in models. Built in model classes >>> >> imply the use of an ORM which then leads you to possibly a SQL-only ORM, >>> >> which then could become tied into every other aspect of the framework. This >>> >> is what killed Django for me, over time. The more that we started using >>> >> other database engines, and things, Django's ORM became such a crutch that >>> >> we decided to look for alternatives - we found Flask. >>> >> >>> >> On Thu, Sep 9, 2010 at 10:31 AM, Davide Muzzarelli >>> >> <d.muzzarelli@dav-muz.net> wrote: >>> >> This presentation talks about Django, but there are things that may be >>> >> useful >>> >> for the evolution of Flask: >>> >> >>> >> http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it >>> >> >>> >> IMHO Flask do it better than Django, and it is written also there. >>> >> >>> >> The suggestion about models is a good thing to think about (I will >>> >> do). >>> >> Actually I fix the configuration of a Flask app moving some function >>> >> calls from >>> >> Flask.__init__(). >>> >> >>> >> What do you think? >>> >> >>> >> Davide Muzzarelli >>> >> www.dav-muz.net >>> >> >>> >> >>> > >> > >
In data giovedì 09 settembre 2010 19:36:00, Tim Conrad ha scritto: > I have nothing against an extension adding this functionality, but I don't > want to *have* to use something. Don't worry Tim, it's just a topic outside the Flask's core. Many developers have similar problems, it's all about community and create good practices that anyone can chose to follow in order to make future (external) applications interoperable. (IMO) We love Flask as is, lean! :) Davide Muzzarelli www.dav-muz.net
I wasn't that concerned about it - I just like that I can make a dumb web app in 10 lines of python that solves stupid problems for me. I've been known to port bash scripts to cgi's so that helpdesk people could monitor something instead of teaching them how to log into a linux box and etc. On Thu, Sep 9, 2010 at 2:08 PM, Davide Muzzarelli <d.muzzarelli@dav-muz.net>wrote: > In data giovedì 09 settembre 2010 19:36:00, Tim Conrad ha scritto: > > I have nothing against an extension adding this functionality, but I > don't > > want to *have* to use something. > > Don't worry Tim, it's just a topic outside the Flask's core. > > Many developers have similar problems, it's all about community and create > good practices that anyone can chose to follow in order to make future > (external) applications interoperable. > > (IMO) We love Flask as is, lean! :) > > Davide Muzzarelli > www.dav-muz.net >
On 9 September 2010 20:54, Tim Conrad <tim@timconrad.org> wrote: > I wasn't that concerned about it - I just like that I can make a dumb web > app in 10 lines of python that solves stupid problems for me. I've been > known to port bash scripts to cgi's so that helpdesk people could monitor > something instead of teaching them how to log into a linux box and etc. That sounds like a very practical niche, where a larger framework would be just overkill.
On 9 September 2010 16:31, Davide Muzzarelli <d.muzzarelli@dav-muz.net> wrote: > This presentation talks about Django, but there are things that may be useful > for the evolution of Flask: > http://www.scribd.com/doc/37113340/Why-Django-Sucks-and-How-we-Can-Fix-it > > IMHO Flask do it better than Django, and it is written also there. > > The suggestion about models is a good thing to think about (I will do). > Actually I fix the configuration of a Flask app moving some function calls from > Flask.__init__(). You should be able to do this (mostly) by subclassing Flask. That's the advantage of an actual class rather than the "magic" setup of Django; you have a tangible instance you can configure, or override as the case may be. > > What do you think? > 1. The main thing I took away from the talk was not the specifics of how to fix Django, rather the difficulty in creating a developer community around a project. On the one hand you want to ensure quality and cohesiveness; on the other you don't want to become a small, close-minded group unwilling to accept new ideas. It's not an easy path to tread. 2. An interesting thought experiment - how easy would it be to create a "Django 2" based on Flask (or Werkzeug, SQLAlchemy, Jinja2 etc) ? Much the same way as Rails 3 is based on Merb ? At the time Django was conceived many of these libraries didn't exist (or were too immature), they were forced into a NIH situation. Given the possibilities now, could you rebuild Django from scratch using existing libraries ?