Forward genetics is the principle of identifying the genetic basis of a phenotype. Reverse genetics is the process of identifying the phenotype coded by a gene of interest (and is much harder and more challenging than forward genetics).
When conducting a forward genetics study, phenotypic screening is used to select for desired characteristics of mutants. This allows the use of genotypic analysis to determine where mutations have knocked out / down the phenotype, and subsequently (via sequencing) allows the identification of where the gene is located.
With reverse genetics, the gene of interest is mutated in vitro before being inserted into a chassis organism. Then, determine if the transcriptome has been altered (by RNA-seq, …) and then find the phenotype that has changed - this part is challenging, due to the breadth of phenotypes present. 96-well plates (BiOLOG) can be used to assay large numbers of phenotypes, determining whether respiration can occur with different sulfur, carbon, … sources. Although this allows many phenotypes to be characterised, it is expensive (due to the specialist plates) and may not identify a phenotype.
With forward genetics, phenotypic screening is done for the desired mutant. This uses genome sequencing and mapping to identify where mutations have impacted the phenotype of interest. Complementation can then be used to determine whether a gene identified through phenotypic screening is wholly responsible for the phenotype of interest - this involves cloning the gene into a plasmid, and introducing it to a mutant cell, determining if the phenotype reverts to wild type.
Bioinformatics tools like BLAST can be used to determine whether the gene of interest has been characterised previously, reducing the amount of work required and simplifying the experiments.