|Date of decision:||20 October 1995|
|Case number:||T 0264/94|
|IPC Class:||G05B 19/42|
|Language of proceedings:||EN|
|Download and more information:||
|Title of application:||A method for preparing numerical control information|
|Applicant name:||FANUC LTD.|
|Opponent name:||Robert Bosch GmbH|
|Relevant legal provisions:||
|Keywords:||Inventive step - yes
State of the art
Novelty - yes
Determination of the technical problem
Need to improve a method
Summary of Facts and Submissions
I. European patent No. 0 044 192 was granted on 13 May 1987 on the basis of European patent application No. 81 303 110.1, filed on 8 July 1981 and claiming priority date of 10 July 1980.
II. An opposition was filed on 4 February 1988 on the grounds that the subject-matter of the patent was not new and did not involve an inventive step (Article 100(a) EPC).
The opponent relied in particular on the document
Spur et al, "Entwicklung einer CNC-Steuerung mit integriertem Programmiersystem für Drehautomaten", ZwF 74(1979) 10, 482-486.
III. On 15 October 1988 the opponent referred additionally to the prior art document
Martin et al. "Ansätze zur Arbeitsbereicherung an NC- Maschinen durch Mikrocomputer", Rationalisierung, 1979- 2, 39-42.
IV. In a communication dated 20 January 1993 the Opposition Division introduced a document on its own motion:
Junike, "Bedienfunktionen einer MPST-Lösung".
This document was said to correspond to a talk held on 23. February 1978 at the conference MPST (Modulares Mehrprozessor Steuersystem). It had initially been filed by the present opponent in a co-pending case relating to similar subject-matter (T 0263/94).
IV. On 11 September 1993 the opponent filed the document
Sauer, "Dialogorientierte Kleinrechner zum Erstellen von NC-Steuerlochstreifen", Maschinenmarkt 85 (1979) 6, 82 - 86.
V. By its decision of 17 January 1994 the Opposition Division maintained the patent in amended form.
VI. Claim 1 as amended reads:
"A method of preparing numerical control information for use in operating a machine tool comprising the steps of:
- programming the shape of an article to be machined;
- displaying characters and/or symbols relating to the shape of the article to be machined; and
- inputting by an operator prescribed data in the form of characters and/or symbols and numerical values relating to the article;
characterised in that:
- sets of representations of respective shapes of articles are prestored as data in a memory by an operator entering via a data input device (1,2,3) a series of items of data ( ,,C,R) which indicate the directions of extent of successive edges of an outline of each respective article, each of said data items ( ,,C,R) being entered separately by the operator and representing a specific direction of extent;
- the prestored shape data is successively recalled for activating a display device (4) according to a predetermined sequence of sets of representations of respective article shapes;
- the article shapes are displayed to an operator as successive inquiries having codes (001, 002, 003...) according to said sequence, by means of characters and/or symbols ( ,,C,R) displaying to the operator the directions of extent of successive edges of the outlines of the displayed article shapes;
- the operator chooses a required shape from amongst the displayed article outlines, by comparing an article shape to be machined with the displayed article outlines as represented by said characters and/or symbols ( ,,C,R) until he locates a displayed article outline which corresponds to the article shape to be machined as regards the directions of extent of the successive edges of its outline;
- the operator inputs the located displayed article shape by using a data input device (1,2,3) to enter the code thereof;
- and subsequently inputs the dimensions of the article shape to be machined by inputting the dimensions to be ascribed to said successive edges of the outline thereof."
VII. On 25 March 1994 the opponent filed a notice of appeal against this decision and paid the prescribed appeal fee. A statement setting out the grounds of appeal was subsequently filed on 29 April 1994.
VIII. In the annex to a summons to attend oral proceedings dated 22 August 1995, the Rapporteur made particular reference to the articles by Junike and by Martin. Oral proceedings were appointed for 19 October 1995.
IX. Oral proceedings were held on 19 and 20 October 1995.
The appellant's (opponent's) arguments in support of his submission that the subject-matter of claim 1 lacked an inventive step can be summarised as follows.
The contested patent was concerned with the preparation of control information for use with a numerically controlled (NC) machine. The technical problem indicated in the description was to "enable even an inexperienced operator to prepare numerical control information relating to arbitrary article shapes quickly, simply and accurately". Such an aim was known from a large number of documents, for example Spur (point 2.1), Sauer (page 83) and Junike (page 105 onwards). In accordance with the method set out in claim 1 of the contested patent this problem was solved basically by the following steps: an operator stored shapes in a memory, the shapes were recalled and displayed on a display device, an operator selected the required shape and entered dimensions which were attributed to the respective edges of the shapes. A similar technique was known from Martin in respect of macros. A macro was a program defining a typical tool movement. It did not specify dimensions values. These were added after a macro had been called, similar to the invention. Martin did not disclose that shapes were displayed; graphical displays were however well known at the relevant date even if their price was prohibitive for many applications. Obviously a graphical display would be desirable for displaying the shapes defined by the macros in Martin. Furthermore, Junike suggested a way of inputting shapes edge by edge using buttons indicating the direction of the edge. A skilled person would consider first using the method described in Junike to produce sets of machine data and then applying the macro principle according to Martin to the sets of data in order to create a shape library. The combination of Martin and Junike thus rendered the invention obvious.
X. The respondent (patentee) argued that a macro, as described in Martin, did not represent a complete shape but merely contained indications as to what the parameters of the shape were. Junike was nearer to the idea of symbolic representations according to the invention; however, it suggested that an edge was always defined together with its dimension value. This was different from the series of data items, excluding dimensions, which the operator stored according to claim 1. Even if Martin perhaps showed the separate treatment of a function and its associated parameter, this was not a feature which could be arbitrarily generalised. It was in particular not applicable to the Junike teaching. Moreover, it had not been shown that the Junike paper was prepublished: it was not certain that the conference at which the paper was presented had been public; there was not even any proof that the lecture corresponding to the Junike paper had taken place.
XI. The appellant requested that the decision under appeal be set aside and the patent be revoked.
XII. The respondent requested that the appeal be dismissed and that the patent be maintained as maintained by the Opposition Division.
Reasons for the Decision
1. The appeal is admissible.
The Board is satisfied that the amendments to the claims made during the opposition proceedings do not extend beyond the content of the application as filed or cause the protection conferred to be extended. The patent therefore complies with Article 123(2),(3) EPC.
3. Novelty not being at issue, only the question of inventive step requires consideration.
4. The invention
The invention as claimed in claim 1 relates to a method of preparing numerical control information for use in operating a machine tool. The salient features of the claim are concerned with the way an operator stores shapes to be machined into a library. The operator enters a series of items which indicate the directions of extent of successive edges of the shape of each article. Dimensions are not stored. Prestored shapes may be displayed on a screen. The directions of extent are displayed by means of symbols such as arrows. In order to prepare machining data for a certain article the operator selects the proper shape and enters, for each edge of the shape, the correct dimension.
5. Prior art
5.1. The Board is of the opinion that the two closest documents on file are the papers by Martin and by Junike.
5.2. The respondent raised the question whether the paper by Junike, said to correspond to a talk held at a conference on 23 February 1978, can be regarded as prior art under Article 54(2) EPC. In view of the following considerations and the conclusions of the Board, cf paragraph 7, this issue need not be resolved by the Board.
5.3. The preamble of claim 1 of the patent can be read onto the teaching of Martin. Furthermore, the method of Martin permits operators of NC machines who do not possess programming skills to create small machining programs. The operator calls a selected one of a plurality of stored "macros". Each macro corresponds to a certain series of tool movements, ie a certain shape. In the single example shown, the shape characteristics do not consist of the direction of successive edges but rather of such conventional parameters as diameters and lengths. These are apparently not ordered in a particular way and would therefore have to be identified, for example from a drawing. The operator inputs, for each parameter, the dimension of the article to be machined.
5.4. Junike discloses, also in the context of permitting a machine operator to create NC programs, a way of defining the shape of an article to be machined using the direction of extent of its successive edges. For each edge, a button symbolising the direction (x-axis, z-axis, diagonal, etc) is activated and the operator is prompted to input the corresponding dimension. There is no mention in the document that items symbolising the direction of edges should be stored separately from the dimension values; on the contrary, the indication of x and z coordinates in the drawings suggests that a complete contour, including dimensions, is stored.
6. The technical problem
Taking Martin as the closest prior art, it can be seen that the rather general technical problem indicated in the description of the contested patent - to "enable even an inexperienced operator to prepare numerical control information relating to arbitrary article shapes quickly, simply and accurately" - has already been solved. The objective problem in this case is, in the opinion of the Board, simply to improve said method further so that the machine operator himself may prepare numerical control information corresponding also to articles whose shape have not been prestored, ie the operator should be able to define not only dimensions but also article outlines.
7. The inventive step
7.1. The macro technique in Martin offers no solution to the aforementioned problem. Stored article shapes can be varied by entering different dimensions, but there is no simple way for an operator to input completely new shapes since the macros have to be programmed. Moreover, the stored shapes cannot be easily visualised since the displayed parameters have to be interpreted by means of a drawing. Also, according to Martin, a creation of a library in the sense as claimed is not possible. Martin does not even disclose how or where the macros are stored.
7.2. Junike offers a solution to part of the problem. Shapes can be entered in very much the same way as according to the invention, using buttons representing different directions of extent of article edges. However, Junike does not hint at the idea that the created shapes could be stored in a library in order to be recalled later. Furthermore, the edges are immediately provided with dimensions, which means that what is finally stored is in principle a complete machining program. When an article is displayed on screen, for example when its dimensions are to be modified, this is apparently done in the form of x and z coordinates, as shown in Figure 4. Such a representation cannot be easily visualised.
7.3. A combination of Martin and Junike does not render the invention obvious either. In Junike, the final form of a shape seems to be a set of coordinates, characterising one single article outline. Martin, however, starts with a number of parameters which define not a single article outline, but a general, dimensionless shape. This is an incompatibility which, in the Board's view, made it difficult for the skilled person to visualise a combination of the teachings. Considering that the processes according both to Martin and to Junike end up with a set of complete numerical control information which can be used directly to control an NC device, the skilled person would probably have seen these teachings as alternatives rather than as complementary.
7.4. It might be added that it would no doubt have been obvious to represent shapes graphically on the screen in the form of contours. Such a method, however, would not correspond to the one claimed in the contested patent. According to claim 1, the article shapes are displayed by means of symbols representing the directions of extent of successive edges. On a fair reading of the claim the symbols must be taken to be identical with the ones previously entered by the operator. Graphical representations in the form of contours are not provided when the stored shapes are displayed.
Thus the available prior art, although disclosing parts of the method, does not suggest such a consistent use of the representation.
7.5. To summarise, the invention solves the problem formulated at point 6 above by a representation of shapes consisting of items which indicate the direction of extent of successive edges of the contour. These items are convenient to use when storing a new shape in a library. Thus the operator himself may create the library and does not need to use macros. Moreover, the same series of items are displayed to the operator when he selects a desired shape to be machined. There is no need for a complete geometrical contour to be displayed, nor for a drawing to serve as an interpretation aid.
7.6. The Board therefore concludes that the subject-matter of claim 1 involves an inventive step.
For these reasons it is decided that:
The appeal is dismissed.