Glass is a fascinating material which has been used for load bearing structures in the past decades. The development of structural glass has been moving towards more transparent structures. The thermal properties of glass can be used to create heat bonds. Heat bonds can be produced in two ways: welding and fusing. With welding, the regions that are to be connected are locally heated. With kiln formed fusing, the elements to be bonded will be globally heated in an oven. This results in the research question: What is the best method, among glass welding and kiln formed glass fusing, to produce heat bonded systems for applications in glass structures? In theory, safety measures consist of three basic strategies: creating secondary load transfer mechanisms, protectioning and overdimensioning. These strategies can be used on the scale of the system and the structure. The heat bonded system could theoretically be laminated, reinforced and tempered. On a structural level, safety can be ensured by a combination of the strategies. To produce glass welds, the glass is preheated in an oven. A small opening in the oven is made and a line burner is used to heat the to be welded area from both sides. One element is pressed down on the other to make the connection. The oven is closed and the specimen is annealed. Finite element models have been used to analyze thermal shock failure during the weld production process. The recommendations for production were to use an additional flame on the bottom of the specimen or to maintain the surroundings at elevated temperature. The latter option is applied in the experiments. The results were three transparent, all-glass, soda lime T-shaped objects. Little residual stress remained in the products. The fused objects were globally heated in an oven. The glass was fused in moulds, made of a gypsum plaster. The products were translucent, all-glass, soda lime, T-shaped objects. The objects deformed significantly during production. Little residual stress remained in the products. Throughout the experiments, factors have changed. An aspect with no direct influence on the product was quenching after maximum temperature. Adapting the mould and polishing the glass edges prior to fusing had an influence on the objects' shape or texture. None of the factors influenced the order of magnitude of the deformation or the seam between the elements. A structural test has been performed to compare the specimens with a glass T-shape bonded with transparent UV-curing adhesive. The connections were loaded in shear. The results of the failure loads and stress at failure are of the same order of magnitude. The largest spread was among the fused specimens, the smallest among the welded. The cause of failure is most probably a combination of elevated stress due to the stiffness and support of the test set-up and peak stress due to irregularities on the glass surface. Other causes could have contributed to failure. The welded glass product requires most time, energy, skill and money. The adhesive product requires least time and energy. The welded and fused objects had neither acceptable dimensional accuracy, nor a smooth fillet in the joint. In the current state of development, glass fusion results in less residual stress, requires less skill and is cheaper to produce. Glass welding results in a transparent glass to glass connection and its structural performance is promising. Further research is encouraged into the heat bonded and adhesive connections to create fully transparent connections for structural glass.