Re: Directory for dictionary databases
On Tue, 28 Apr 1998, Bob Hilliard wrote: > I don't think a mono-lingual dictionary is "translation related" > data. Hmmm, that's a nice topic for discussion. For me as a German native speaker ar monolingual dictionaries like wordnet sometimes a more helpful tool than a bi-lingual dictionary. It depends on the knowledge of the translator. So if I don't know an English word I try wordnet first and if I can't imagine the meaning then the second try is a bi-lingual wordlist. In this way I think it *is* translation related. > I agree that bi-lingual files deserve a separate location, but > a structure such as you suggest would only be useful if accepted as a > standard. So far, at least, the FHS committee has avoided specifying in > much detail the subdirectory structure (/usr/share/man and > usr/share/dict are exceptions). They do recommend : > > | ... > > We should keep in mind that databases for programs like dictd (I > don't know if this is true of WordNet) are formatted so that it is > difficult, at best, to access them with other programs, so there is > less need to keep them together. Wordnet comes with a library for accessing the database. So you could write your own programs which handle the data. > I have no contact with the FHS committee, and I don't have time > to get involved in it either. There is an FHS mailing list, but I > don't know its address. The current draft of the FHS is available at > tsx-11.mit.edu:/pub/linux/docs/linux-standards/fsstnd. As I said I have no time to do this, too. May be someone of the other developers could say if he sees the same problem with directory structure like me and if there is any suggestion how to cope with. I will definitely not involve in FHS development. I would prefer to collect the "translation related data" (in the sense I mentioned above) in one directory under /usr/share but if I'm the only one who thinks so the wordnet database will end in /usr/share/wordnet. Regards Andreas. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Directory for dictionary databases
Andreas Tille <[EMAIL PROTECTED]> writes: > But in my opinion it is a bad solution if every "translation related" > data should be stored in different directories. I think a directory > structure should be as logical as possible. Im not familiar with > FHS and havn't enough time to learn about it. But in my opinion it > is necessary to store data together if they have the same purpose. > I would suppose to store translation related data in > > /usr/share/lang > or > /usr/share/trans [-lation] > or > /usr/share/lingua > or may be > /usr/share/lingua-mono AND > /usr/share/lingua-bi (lingua-multi) > > The latter case means to store monolingual databases like wordnet and > dict in lingua-mono and bilingual databases (I plan to maintain a > English-German dictionary) in lingua-bi. (We should think about > the Euro-wordnet project which is multilingual.) > Supporting this hierarchie wie can store in this directory single > file dictionaries and in a subdirectory if there are more files like > in the case of wordnet. I don't think a mono-lingual dictionary is "translation related" data. I agree that bi-lingual files deserve a separate location, but a structure such as you suggest would only be useful if accepted as a standard. So far, at least, the FHS committee has avoided specifying in much detail the subdirectory structure (/usr/share/man and usr/share/dict are exceptions). They do recommend : | Any program or package which contains or requires data that doesn't need | to be modified should store that data in /usr/share (or | /usr/local/share, if installed locally). It is recommended that a | subdirectory be used in /usr/share for this purpose. We should keep in mind that databases for programs like dictd (I don't know if this is true of WordNet) are formatted so that it is difficult, at best, to access them with other programs, so there is less need to keep them together. > I think if we don't do that we will end up in so many directories in > /usr/share after some years that it would be horror. Think of NLS > and such stuff. There will be more and more needs for dictionaries > and we should do it right from the beginning. I don't see this as too much of a problem. Look at /usr/doc, with a subdirectory for every executable. > Please foreward this to the FHS team because I'm not able to involve > in this theme. I have no contact with the FHS committee, and I don't have time to get involved in it either. There is an FHS mailing list, but I don't know its address. The current draft of the FHS is available at tsx-11.mit.edu:/pub/linux/docs/linux-standards/fsstnd. Bob -- _ |_) _ |_ Robert D. Hilliard<[EMAIL PROTECTED]> |_) (_) |_) Palm City, FL USAPGP Key ID: A8E40EB9 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Directory for dictionary databases
On Mon, 27 Apr 1998, Bob Hilliard wrote: > In earlier discussions, it was agreed that the proper location > for dictionary databases under the FHS would be > /usr/share/dict/. > > Subsequently, I have obtained a copy of the FHS, and have > discussed this subject with the upstream author of the dictd package. > > An abbreviated extract from the FHS follows is appended to this > message. > > It appears that the FHS considers the /usr/share/dict directory > to be for word lists, such as those used for spell checkers, not for > dictionaries in the wider sense. > > Rik Faith <[EMAIL PROTECTED]>, the author of dictd, says: > > After reviewing the FHS, I agree with Rik, and will build the > dictionary packages that I expect to upload tomorrow to install in > /usr/share/dictd. > > I suggest that you consider /usr/share/wordnet for your packages. ... > 4.8 /usr/share : Architecture-independent data > > /usr/share -- Architecture-independent data > | > +-dict Word lists > . . . . . > > Any program or package which contains or requires data that doesn't need > to be modified should store that data in /usr/share (or > /usr/local/share, if installed locally). It is recommended that a > subdirectory be used in /usr/share for this purpose. > > . . . . . > 4.8.1 /usr/share/dict : Word lists > > Recommended files for /usr/share/dict: > > { words } > > Traditionally this directory contains only the English words file, which > is used by look(1) and various spelling programs. > . . . . . > Word lists for other languages may be added using the English name for > that language, e.g., /usr/share/dict/french > . . . . . > > Other word lists, such as the web2 "dictionary" should be included here, > if present. > > BEGIN RATIONALE > > The reason that only word lists are located here is that they are the > only files common to all spell checkers. > > END RATIONALE OK, that seems to be a good reason to store only spell check related data in /usr/share/dict. But in my opinion it is a bad solution if every "translation related" data should be stored in different directories. I think a directory structure should be as logical as possible. Im not familiar with FHS and havn't enough time to learn about it. But in my opinion it is necessary to store data together if they have the same purpose. I would suppose to store translation related data in /usr/share/lang or /usr/share/trans [-lation] or /usr/share/lingua or may be /usr/share/lingua-mono AND /usr/share/lingua-bi (lingua-multi) The latter case means to store monolingual databases like wordnet and dict in lingua-mono and bilingual databases (I plan to maintain a English-German dictionary) in lingua-bi. (We should think about the Euro-wordnet project which is multilingual.) Supporting this hierarchie wie can store in this directory single file dictionaries and in a subdirectory if there are more files like in the case of wordnet. I think if we don't do that we will end up in so many directories in /usr/share after some years that it would be horror. Think of NLS and such stuff. There will be more and more needs for dictionaries and we should do it right from the beginning. Please foreward this to the FHS team because I'm not able to involve in this theme. Just my two cents Andreas. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Directory for dictionary databases
In earlier discussions, it was agreed that the proper location for dictionary databases under the FHS would be /usr/share/dict/. Subsequently, I have obtained a copy of the FHS, and have discussed this subject with the upstream author of the dictd package. An abbreviated extract from the FHS follows is appended to this message. It appears that the FHS considers the /usr/share/dict directory to be for word lists, such as those used for spell checkers, not for dictionaries in the wider sense. Rik Faith <[EMAIL PROTECTED]>, the author of dictd, says: > I argue that dictd is an application which stores read only information in > a proprietary[1] format which is not easily used by other applications, and, > as such, the data should be stored in the top level directory: > /usr/share/dictd After reviewing the FHS, I agree with Rik, and will build the dictionary packages that I expect to upload tomorrow to install in /usr/share/dictd. I suggest that you consider /usr/share/wordnet for your packages. Bob [1] This does not mean proprietary in the non-free sense. Rik means unique to the dictd package. -- _ |_) _ |_ Robert D. Hilliard<[EMAIL PROTECTED]> |_) (_) |_) Palm City, FL USAPGP Key ID: A8E40EB9 Abbreviated extract from the Filesystem Hierarchy Standard -- Version 2.0 4.8 /usr/share : Architecture-independent data /usr/share -- Architecture-independent data | +-dict Word lists . . . . . Any program or package which contains or requires data that doesn't need to be modified should store that data in /usr/share (or /usr/local/share, if installed locally). It is recommended that a subdirectory be used in /usr/share for this purpose. . . . . . 4.8.1 /usr/share/dict : Word lists Recommended files for /usr/share/dict: { words } Traditionally this directory contains only the English words file, which is used by look(1) and various spelling programs. . . . . . Word lists for other languages may be added using the English name for that language, e.g., /usr/share/dict/french . . . . . Other word lists, such as the web2 "dictionary" should be included here, if present. BEGIN RATIONALE The reason that only word lists are located here is that they are the only files common to all spell checkers. END RATIONALE -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]