Data validation dialect

Hi, I'm now ready to begin to experiment some minimal GUI testing and data validation for a small project of my own. I already looked at an interesting proposed way to dynamically create and validate input screens - in a way that permits some reuse of initial declarations. (see Graham's work at http://en.wikibooks.org/wiki/REBOL_Programming/Language_Features/Parse/Screen_Validation) But for the most cumbersome and repetitive part, that is data validation I didn't find any dialect DSL using REBOL to simplify the complete specs definition part of the task. However even if I plan to begin writing one myself, I would like if any of you already did some work in this domain. Also found some work in old rebol forces magazine about a data generator dialect (from Joel Neely) that could be of some help (http://www.rebolforces.com/zine/rzine-1-02/) In the mean time, I found some links related to the subject of data validation using DSL (dialects) that I would like to share with those of you that have some interest relating to the subject : Why current validation methods often fail (security issues) - where and how to do it properly ================================================================== http://research.corsaire.com/whitepapers/060116-a-modular-approach-to-data-validation.pdf Some more thoughts : ================ http://c2.com/cgi/wiki?SeparateDomainFromPresentation Properties- model ============= http://parasol.cs.tamu.edu/~jarvi/papers/gpce08.pdf Issues related to User-interfaces (related to the Olivanova method) ======================== http://www.upgrade-cepis.org/issues/2003/1/up4-1Molina.pdf A form data validation DSL using Ajax and JS ================================= http://ojay.othermedia.org/articles/forms.html Regards, Gerard

Gerard I'm sure that script can be readily modified to include ranges of values for the field validation. On Fri, Dec 18, 2009 at 6:39 PM, G=E9rard C=F4t=E9 <gerardcote-gmail.com> wrote:

> Hi, > I'm now ready to begin to experiment some minimal GUI testing and data > and how to do it properly > =========================

> http://research.corsaire.com/whitepapers/060116-a-modular-approach-to-data-validation.pdf > Some more thoughts : > A form data validation DSL using Ajax and JS > =========================

> http://ojay.othermedia.org/articles/forms.html > > Regards, > Gerard > -- > To unsubscribe from the list, just send an email to > lists at rebol.com with unsubscribe as the subject. >

Hi Graham, This is what I will begin with but my true goal is to have both an implementation and a dialect for describing what the data validation rules specs are, be it localized in the app itself or in an external shared data object or centralized dictionary. Kind of minimalist general business rule encoding - like is done by databases systems and centralized data-centric objects in a true multi-tier app - eventually I'd like to have such a system running ... since I gradually want to move the business rules from the presentation layer (GUI) to a common data layer, which is somewaht easier to share across a distributed app, via services, for example. For the moment, I'm trying to figure out most of the common ways in use that describe form (and field) data specs and validation rules and what is better done at the UI level. Sometimes but not always, this kind of validation is driven by the data type and the domain of values each type permits. So I want to collect some of those ways that are in actual use, just to see if there is some general logic that can be applied through my own "data validation definition dialect" design. Here is my first summary : (sure there are others and would be glad it anoybody would enrich/correct this first draft) (My only past experience in this domain comes from the design of VB input forms, thus validating at the field and at the form levels. Everything else comes from recent readings. So many misconceptions/missing info are more than probable). At the field level we generally find those relevant attributes, independently of the GUI system, I think : ---------------------------------------------------------------------------------------------------------------------------------------------------------- - Mandatory/optional - Protected/writable - Visible/hidden - Size, position, color, etc... - Behavior (local event management and simple validation rules - most related to datatype, i.e. Numeric! - discrete or continuous : min. max, interval(s), inclusion or exclusion related boundary conditions expressed by means of functions, etc...; String! : pattern recognition, uppercase, phone number, valid date, ...) At the form level we have to take into account for : ------------------------------------------------------------------------ - Behavior (Global event management and dependencies between some or all of the fields - expressed by more elaborate validation/derivation rules) Then there are also what are called the other more general business rules - generally taken into acount at the object/database level -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- - Behavior/dependencies related to the business way of doing things (policies) Just asked myself about what "formal or natural - dialect" language could be used for defining the specs of these validation/business rules AND WHERE TO PUT /when to use THEM ? ==================================================================================================================================== Began some research and found many alternatives already in use : 1- saw an article about the OCL part of UML that is used playing this role for objects classes constraints (wikipedia) 2- saw in one of my old books a reference to BNF stating that by 1984 it was the formal tool to be used for string pattern recognition (Here we have parse coupled with PEGs) 3- OASIS ia another formal description language used as the basis for the OlivaNova OO-Method and the Just-UI (See URLs : www.care-t.com/index.asp www.ercim.org/publication/Ercim_News/enw57/pastor.html www.dsic.upv.es/~einsfran/papers/06-CAISE97.PDF portal.acm.org/citation.cfm?id=964442.964522 www.dsic.upv.es/grupos/oom/oasis.html - this last one the OSAIS 3.0 spec is available in Spanish only) 4- Finally in the prolific Java/Ruby World there are descriptive testing tools (Fit, JBehave and RSpec) that are more of the By example kind - They describe the same thing the xUnit family members are able to test but using a behavior related vocabulary The refs about the tools related to BDD from Dan North and David Chelimsky - (URL dannorth.net/introducing-bdd) See also Agiledox/TestDox tools (URL agiledox.sourceforge.net/) 5- Then another view desserves a good mention about describing and implementing business rules in REBOL itself - (URL: computer-programming-languages.suite101.com/article.cfm/how_to_createa_rebol_dialect#ixzz0KnhL2yyi&C) Why didn't I think about it before ? probably because I want to get a descriptive approach and I wil have to create some dialect to do so. 6 Also found this about the new VB 9.0 Query mode - Will try to get my hands for creating a similar one for REBOL when ready to do so (URL : tinyurl.com/z8z82)

Hi Graham, I'm revising the doc written by Henrik for his VDI Extensions, and he already created a Validate add-on to the VID dialect for helping with face validation. See his user manual for more info. Great documentation work Henrik. As I was particularly interested by the Style creation under VID/View, I've read your doc about this aspect and while not easy to grasp every detail when not already familiar with the VID/View internals, it's very comprehensible and will be useful for anyone interested by this level of expertise, like me. Really Top Notch addition to the other View/VID documentation. There was a real need for this doc. Hope to be able to do something real with it but for the moment, I have to help you more with the testing of all these existing new styles and functionalities you added to VID. After a long search about automated GUI testing, using Ruby, this was of no help to me since I don't know how to simulate an event for View to accept it as if it was really done by the user. I can get the contents of a real event datatype! but can't interpret those and so can't create one myself, ignoring what to send to View as an event ... ans how to insert it into the event queue. May be later I will be better at understanding all of these concepts. For now I'll limit myself at describing the behaviors of my tests (a bit like what is done with Rspec and JBehave) but theese specs will not be executables! They'll serve only for reference doc. But first I have to read more of your doc to better know what are the new tools you added for developers. And no Maxim, I didn't forget you and GLayout. I'll test it in parallel if this is not too much for me, or in alternance. That's a lot to read, experiment with and understand but this is what I want - learning REBOL to be able to use it in a decent manner !!! Thanks to all past, actual and future contributors. Newcomers need you and in their name I thank you all again. Enjoy this great OFF time of the year - if you are eligible to do so ! -- G=E9rard C=F4t=E9 Qu=E9bec, Canada