Guest lecture by Prof. Leis from TUM

Prof. Leis began his talk in the special lecture series of the Elite Graduate Program “Software Engineering” by dis­cuss­ing the rela­tional data model and its great suc­cess over the past dec­ades. Close­ly asso­ciated with the rela­tional data model is the query lan­guage SQL, which - due to the suc­cess of the rela­tional mod­el - also be­came very popu­lar, de­spite its obvi­ous short­com­ings. 

One of SQL's weaknesses is its highly incon­sistent syn­tax across dif­ferent data­base sys­tems. Alt­hough there is a standard, the SQL capa­bili­ties of data­base man­age­ment sys­tems (DBMS) vary wide­ly. Hard­ly any DBMS sup­ports the full standard, and al­most every sys­tem intro­duces its own lan­guage exten­sions. This re­duces users’ flexi­bility and cre­ates ven­dor lock-in. Fur­ther­more, SQL is very ver­bose and prone to errors in prac­tice. 

In addi­tion to syn­tactic issues, the se­man­tics of SQL que­ries is not clear­ly de­fined. The same SQL query can pro­duce dif­ferent results on dif­ferent DBMS. This signif­icant­ly hin­ders the tech­nolo­gy-ag­nostic de­velopment of soft­ware sys­tems that rely on data­bases.

Prof. Leis then pre­sent­ed two re­search pro­jects aimed at de­velop­ing a better query lan­guage while pre­serv­ing the suc­cess­ful rela­tional data mod­el: SaneQL and SaneIR. SaneQL is a new query lan­guage with a mod­ern, com­pact syn­tax that elimi­nates SQL’s syn­tactic clut­ter, mak­ing it signif­icant­ly easier for users. To define a con­sistent se­man­tics, SaneIR serves as an inter­medi­ate repre­senta­tion—a kind of lingua franca be­tween data­base sys­tems—in­tend­ed to sup­port com­pati­bility and porta­bility.

Text:  Dr. Dominik Haneberg, Elite Graduate Program “Software Engineering”