Working on a building design to examine the options for balancing heat gain/ loss / lighting requires suitable:.

• typical sky conditions in the location and appropriate climatic data (LINK)(LINK)
• appropriate energy use assessment tools to integrate with the lighting tool(s) (LINK)
• appropriate models for user behaviour (LINK)
• validation procedures,
• a data base of glazing materials

rules of thumb
Traditional rules of thumb widespread in text books and used in industry sustainability certification processes like LEED are singularly unhelpful in this regard. The consensus of these various expressions of daylight rules of thumb seems to be that if you have clear glazing and windows unobstructed inside and out then adequate daylight for many office applications is achievable in temperate climates up to 1.5-2 times the head height of the window back in from the window. A paper presented at the Berkeley workshop for this IEA Research Task confirmed that this guidance seems to be about right in terms of light quantity.
If the window is obstructed by blinds say, then the depth to which good daylight can be achieved falls to only 1x the head height of the windows. As with all rules of thumb a lot goes unsaid. What is normally assumed here is a strip window the width of the room; it is also assumed that if the sill is below the work surface the light levels will not be affected. What is problematic is that this is intended to be the starting point for good design, not the end point. If this is all the information that is used to “design” daylit buildings then the risk is high that the building will be experience glare, overheating and excessive heat loss, creating an uncomfortable, low productivity environment that no amount of extra energy expended on lighting, cooling and heating will fully solve.
However, the clear message of this re-analysis of this rule of thumb is that daylighting is only possible along the perimeter edge of a deep plan building, or in buildings that have a short distance from the perimeter to the core.
state of the art in daylight design tools
In the special issue of Energy and Buildings Reinhart and Fitz "provide a snapshot of the current use of daylight simulations during building design based on an online survey of 185 individuals from 27 countries. The paper confirms the existence of a growing group of designers, engineers, and consultants who routinely use daylight simulations during building design. It further provides an overview of which tools these individuals use, how they use them, and why."

data on advanced light redirecting materials
Validation studies have shown that lighting conditions can be simulated "in daylit buildings featuring standard facade elements such as glazings, lightshelves, and diffusing blinds. On the other hand, modeling more advanced, light re-directing complex fenestration systems(CFS), such as laser cut panels or specular blinds, remains a challenge even for the most sophisticated tools".
Bidirectional Transmittance/ Reflectance Distribution functions (BTDF/BRDF) are used to mathematically represent the interaction of incoming light with a CFS. Andersen and de Boer discuss the benefits and shortcoming of existing and planned experimental and numerical tools to determine the BTDF and/or BRDF of a variety of CFSs. de Boer separately describes how to combine BTDF data from an arbitrary CFS with external sky conditions in order to determine indoor illuminance conditions. The method has been implemented in a new simulation tool that also features a BTDF database.
data on the algorithms that can be used to represent advanced glazing systems
One of the major concerns of the participants in Task 31 has been ensuring simulation of building properties in design studies can be as accurate as possible. This has meant a systematic documentation of the algorithms like BTDF’s that exist for the characterisation of the interactions of light and buildings or building materials. A library of trusted formulae that can be used by lighting simulation computer program developers has been compiled. LINK Subtask C.
performance prediction requires accurate models of the sky relevant to local climates
In daylight, a common problem shared by physical models and digital models is the definition of a sky that is representative of all the skies that a proposed building will experience in even a year of operation. As noted earlier, new techniques have been developed that permit lighting simulations to be completed for every daylit hour of a standard year such as happens in thermal design. For those not able yet to use these techniques, and again in conjunction with the CIE, a standard set of sky descriptions has been developed.
LIGHTSWITCH - a model of user behaviour
Early within this IEA Research Task a model of user behaviour called ‘Lightswitch’ was proposed. It describes occupant use of lighting and shading devices in offices. Bourgeois, Reinhart, and Macdonald introduced "occupant behavior modeling into the whole-building simulation program ESP-r. Example simulations of the enhanced ESP-r version demonstrate that, depending on the occupant behavior model chosen, primary energy use for heating, lighting, and cooling in a sidelit office can vary by more 40%."
validation / calibration tests of the accuracy of light rendering software
Tests have been developed in collaboration with the CIE for the validation of computer simulation software. A paper by Maamari, Fontoynont, and Adra within the special issue of Energy and Buildings describes these tests which are for both daylight and artificial light simulations. The paper tests two lighting computer programs: Lightscape 3.2 and Relux Professional 2004. The intention is that this becomes a set of calibration tests for lighting software that a0 can be used by software developers in the development and diagnosis of the successful operation of their software; and b) can be established as the minimum standard which a computer simulation must reach in order to be used for code compliance checking. The full report on the tests is available from the CIE. The results of the application of these tests to some common software packages has been summarised on a separate page in this web site.
Reinhart and Selkowitz note: From Maamari, Andersen, de Boer, Carroll, Dumortier, and Greenup compare the ability of four simulation programs to reproduce indoor illuminances through a CFS measured in a test box, illuminated under either real overcast sky conditions or an artificial sky. The CFSs investigated are a laser cut panel and a crenellated plastic panel. The investigated programs generally modeled indoor illuminances with an accuracy of around 20% under overcast sky conditions. This number is comparable to earlier RADIANCE validation studies that involved standard double glazings. Reinhart and Andersen carried out a validation study for a translucent panel. The optical properties of the panel were determined via goniophotometer and integrating sphere measurements and were converted into a RADIANCE material model. A comparison of measurements under varying sky conditions in a full-scale mockup that featured the translucent panel and RADIANCE simulations revealed that RADIANCE could reproduce the measured indoor illuminances with an accuracy of 20%. The paper further demonstrates that this type of simulation error is sufficient for practical design applications such as daylight factor and daylight autonomy calculations, or predictions of annual electric lighting use.