Instructions for Anonymous FTP of the Alvey Natural Language Tools ------------------------------------------------------------------ (Release 4) from Cambridge University Computer Laboratory --------------------------------------------------------- The Alvey Natural Language Tools (ANLT) are held in the directory `nltools' on the FTP server ftp.cl.cam.ac.uk (128.232.0.56). The files containing grammars, lexicons, source code and full system documentation are encrypted; to extract these you will need to use a machine running Unix, and also to have been issued with a decryption key. However, technical reports describing the ANLT are available in non-encrypted form; you may freely fetch these for examination. Fetching the Reports -------------------- To fetch the reports use anonymous FTP (login with user name anonymous, and password your e-mail address), go to the directory `nltools/reports', get a list of the files in it, and fetch the ones you are interested in, or all of them. The file `README' says what each of the files contains. You could do this by: $ ftp ftp.cl.cam.ac.uk Connected to swan.cl.cam.ac.uk. 220- swan.cl.cam.ac.uk FTP server (Version 5.60+UA) ready. ... Name (ftp.cl.cam.ac.uk:jac): anonymous Password (ftp.cl.cam.ac.uk:anonymous): ... ftp> cd nltools/reports 250 CWD command successful. ftp> ls ... README gde.ps grammar.ps morph.plain ... ftp> get gde.ps ... ftp> get morph.plain ... ftp> quit 221 Goodbye. (The $ is the Unix shell command prompt). If the FTP command does not know about the address ftp.cl.cam.ac.uk, try giving the command the internet number (128.232.0.56) instead. The file 'morph.plain' is ASCII text and so may be viewed or printed out just as it is. The files whose names end in '.ps' are Postscript. The full ANLT system contains the original (LaTeX format etc.) versions of these documents. Fetching the Full ANLT System ----------------------------- To fetch the full ANLT system from the server use anonymous FTP (login with user name anonymous, and password your e-mail address), go to the directory `nltools' and, in `binary mode', fetch three files: 1. the program to decrypt the ANLT; depending on the type of machine you are using, one of decrypt.sun3 for Sun 3 decrypt.sun4 for Sun 4 or Sun Sparc decrypt.hpux for Hewlett Packard 9000/3nn running HP-UX decrypt.mips.ultrix for DECStation 3000 or 5000 series running Ultrix decrypt.tar for any other type of Unix machine, 2. some encrypted data to check that you have decryption working properly; depending on the type of machine you are using when you decrypt it, either test.l for `little-endian' machines (DEC Vax, MicroVax, DECStation), or test.b for `big-endian' machines (virtually all other machines including Sun 3/4/Sparc, Hewlett Packard, Sequent, IBM RISC, 80386), 3. the ANLT itself; depending on the type of machine you are using when you decrypt it, either anlt4.tar.Zl for `little-endian' machines anlt4.tar.Zb for `big-endian' machines IT IS VITAL that you fetch all of these files in binary mode. For example, on a Sun 3 (a `big-endian' machine) you would do the following: $ ftp ftp.cl.cam.ac.uk Connected to swan.cl.cam.ac.uk. 220- swan.cl.cam.ac.uk FTP server (Version 5.60+UA) ready. ... Name (ftp.cl.cam.ac.uk:jac): anonymous Password (ftp.cl.cam.ac.uk:anonymous): ... ftp> cd nltools 250 CWD command successful. ftp> binary 200 Type set to I. ftp> get decrypt.sun3 ... ftp> get test.b ... ftp> get anlt4.tar.Zb ... ftp> quit 221 Goodbye. Make sure that the three files you have just fetched are in the disc directory that you want the ANLT to finally end up in, and that this is your current directory. If not, move them there and `cd' to it. If you fetched the file `decrypt.tar' (because your machine is not one of the four types for which the decryption program is already set up), then you will have to `untar' the file and compile the program in it using the Unix `make' command. When this is done, the decryption program will be in a file called `decrypt': $ tar xvf decrypt.tar ... $ make ... You will now have the decryption program in an executable form (in a file called one of `decrypt.sun3', `decrypt.sun4', `decrypt.hpux', `decrypt.mips.ultrix' or `decrypt'). Next, mark it as actually executable, and run it to check that it is working properly using the encrypted test data file, quoting the decryption key that was issued to you. The decryption key is case-sensitive so you must type the upper- and lower-case letters and numbers EXACTLY as they were given. $ chmod +x decrypt.sun3 $ decrypt.sun3 test.b test Passwd: Verify: $ cat test ... The contents of the file `test' which are displayed should be English text. If the decryption program gave an error then it is likely that you fetched a version of it that is unsuitable for your machine: either fetch the correct one, or fetch the `decrypt.tar' file and compile your own version of it. If the decryption program ran, but the text in the `test' file is not readable, then you probably fetched the wrong `-endian' version of the decrypted files for your machine, or typed the decryption key incorrectly. Try decrypting the test data again, typing the decryption key carefully; if still unsuccessful fetch the opposite `-endian' files and try them. The ANLT itself is contained in an encrypted compressed `tar' archive. To extract the files in it, decrypt the file quoting the decryption key, uncompress the result of that, and finally `untar' it: $ decrypt.sun3 anlt4.tar.Zb anlt4.tar.Z Passwd: Verify: $ uncompress -c anlt4.tar.Z | tar xvf - This process will require approximately 13 megabytes of free disc space. The decryption step may take up to an hour on a Sun 3 (less on a faster type of machine), the uncompress and `untar' step a few minutes. Once this is done you can delete the ANLT archive files and the decryption program and test data: $ rm anlt4* decrypt* test* You will now have all the files that comprise the ANLT in your current directory. They will occupy about 10 megabytes on disc. You may want to make yourself the owner of the files; ask a Unix expert locally to do this for you if you don't know how to. For instructions on what to do next and on how to install the ANLT system, refer to the file called README that will have just been created.