In computing, a hardware code page (HWCP) refers to a code page supported natively by a hardware device such as a display adapter or printer. The glyphs to present the characters are stored in the alphanumeric character generator's resident read-only memory (like ROM or flash) and are thus not user-changeable.[1][2][3] They are available for use by the system without having to load any font definitions into the device first. Startup messages issued by a PC's System BIOS or displayed by an operating system before initializing its own code page switching logic and font management and before switching to graphics mode are displayed in a computer's default hardware code page.
In North American IBM-compatible PCs, the hardware code page of the display adapter is typically code page 437.[4] However, various portable machines[5][6][7] as well as (Eastern) European, Arabic, Middle Eastern and Asian PCs used a number of other code pages as their hardware code page,[1] including code page 100 ("Hebrew"),[8] 151 ("Nafitha Arabic"),[9] 667 ("Mazovia"),[1][10] 737 ("Greek"), 850 ("Multilingual"),[7][nb 1] encodings like "Roman-8",[5][6] "Kamenický",[1][10][11] "KOI-8", "MIK",[10] and others. Most display adapters support a single 8-bit hardware code page only.[1] The bitmaps were often stored in an EPROM[12][13][4] in a DIP socket.[4] At most, the hardware code page to be activated was user-selectable via jumpers,[14] configuration EEPROMs[5][6] or CMOS setup.[15] However, some of the display adapters designed for Eastern European, Arabic and Hebrew PCs supported multiple software-switchable hardware code pages, also named font pages,[1] selectable via I/O ports[4] or additional BIOS functions.[12][13]
In contrast to this, printers frequently support several user-switchable character sets, often including various variants of the 7-bit ISO/IEC 646 character sets such as code page 367 ("ISO/IEC 646-US / ASCII"[1]), sometimes also a couple of 8-bit code pages like code page 437,[10][16][17] 850,[10][16][17] 851,[10] 852,[10][18] 853,[19] 855,[10][18] 857,[19] 860,[10][16][17] 861,[17] 863,[10][16][17] 865,[10][16][17] and 866.[10][18][20][21] Printers for the Eastern European or Middle Eastern markets sometimes support other locale-specific hardware code pages to choose from. They can be selected via DIP switches or configuration menus on the printer, or via specific escape sequences.[1][nb 2]
When operating systems initialize their code page switching logic, they need to know but have no means to determine the previously active hardware code page by themselves. Therefore, for code page switching to work correctly, the hardware code page needs to be specified.[1][22][23][24]
Under DOS and Windows 9x this is accomplished by specifying the hardware code page as a parameter (hwcp) to the device drivers DISPLAY.SYS and PRINTER.SYS in CONFIG.SYS:
DEVICE=…\DISPLAY.SYS CON=(type,hwcp,n|(n,m))[1][2][22][23][24][25][26]DEVICE=…\PRINTER.SYS PRN=(type,hwcp,n)[1][22][23][24][25][26]
If multiple hardware code pages are supported in OEM issues, the first hardware code page (hwcp1) in the list specifies the default hardware code page:
DEVICE=…\DISPLAY.SYS CON=(type,(hwcp1,hwcp2,…),n|(n,m))[9]DEVICE=…\PRINTER.SYS PRN=(type,(hwcp1,hwcp2,…),n)[9]
If no hardware code page(s) are specified, these drivers default either to a dummy code page number 999[1][23][24] or assume the hardware code page to be equal to the primary code page (the first code page listed in COUNTRY.SYS files for a particular country[27] with the country code either specified in the CONFIG.SYS COUNTRY directive or assumed to be the operating system's internal default, usually 1 (US) in Western issues of DOS).[1]
In many English-speaking countries, the primary code page is either 437 (f.e. in the US) or 850 (f.e. in the UK, Ireland and Canada),[23][24] so that, without specifying a different code page, the system would often assume one of these to be the corresponding device's default hardware code page as well.[1]
If a hardware code page does not match one of those with official code page assignments, an arbitrary number from the range 57344–61439 (E000h–EFFFh) for user-definable code pages or 65280–65533 (FF00h–FFFDh) for private use code pages could be specified per IBM CDRA to give the operating system a non-conflictive "handle" to select that code page.
Arabic and Hebrew MS-DOS do not use DISPLAY.SYS and PRINTER.SYS, but provide similar facilities using ARABIC.COM, HEBREW.COM, and SK_HGC.COM.[9]
Hardware code pages are also OEM code pages. The designation "OEM", for "original equipment manufacturer", indicates that the character set could be changed by the manufacturer to meet different markets.[28][29]
However, OEM code pages do not necessarily reside in ROM, but include so called prepared code pages,[2][25] (aka downloadable character sets or downloadable fonts), character sets loaded as raster fonts into the font RAM of suitable display adapters (like Sirius 1/Victor 9000,[30] NEC APC,[31] HP 100LX/200LX/700LX,[15] Persyst's BoB Color Adapter,[32] Hercules' HGC+,[33][13][34] InColor[33][35] and Network Plus[36][37] with RAMFONT, and IBM's MCGA,[33][38][13] EGA,[38][13] VGA,[38][13] etc.) and printers as well.[39][40][41][42][43] Hence, the group of OEM code pages is a superset of hardware code pages.
On Epson, NEC and Fujitsu ESC/P compatible printers, the escape sequence to switch to various hardware code pages is ESC R (n) with (n) = 0..254 being a code associated with specific character sets. The range of supported codes depends on the printer model and installed font options.
Paul, Matthias R. (2001-08-15). "Changing codepages in FreeDOS" (Technical design specification). Archived from the original on 2016-06-06. Retrieved 2016-06-06. […] The […] active system codepage […] sometimes also called global codepage […] is the codepage currently in use for the DOS kernel, for example the collating order and other I18N tables depend on this codepage. While in most cases the system codepage reflects the codepage used for devices such as console (DISPLAY.SYS), printers (PRINTER.SYS), and keyboard (KEYB.COM), this is not a requirement as there are means to switch the individual device codepages without changing the system codepage (for example using MODE). Further, in dealing with NLS issues, it is important to distinguish between the currently active, the hardware, and the primary codepages for any device and/or the system. The hardware codepage is the default codepage supported by a device after initialization of the hardware. It normally defaults to 437 for the console, but some European clones have 850 stored in the video font ROMs, and I have been told that some Czech CGA and MDA cards had Kamenicky codepage 867 (aka Wordperfect codepage 895) stored in their ROMs. Most printers default to 367 (ISO-646-US 7-bit) or 437 on power up, and the default system, or global, codepage for the system is always 437 under DR-DOS […] but under MS-DOS/PC DOS it reflects the codepage entry in COUNTRY=country,codepage,filespec where it must be one of the codepages valid for the given country as per COUNTRY.SYS filespec - this is 437 (or 850 on newer systems) in most cases, but not necessarily. The default country is 1 (USA), the default filespec is "\COUNTRY.SYS", and if no filespec is found the system will work with the statically pre-initialized data in the kernel which is the data for the USA in most cases. If no codepage is given with COUNTRY, the DOS BIOS will search for the first codepage entry for a given country (or 1), this is called the primary codepage. The codepage value of 0 is used internally to retrieve this codepage, and for the COUNTRY.SYS file parsers in the DOS BIOS and NLSFUNC, a codepage value of zero means "match any codepage, whatever comes first for the given country code". So, it depends on the order of codepage entries in the COUNTRY.SYS file if the system will default to 437, 850, or something else. […] However, some Arabic and Hebrew HGC and CGA cards have multiple hardware fonts, so called fontpages, […] so that a device may support multiple hardware codepages at the same time. The corresponding fontpage functions are INT 2Fh/AD41h and INT 2Fh/AD42h. […] COUNTRY.SYS and KEYBOARD.SYS contain only two codepage entries for a given country in Western issues of DOS. (In Arabic and Hebrew issues there can be up to 8 codepages for one country […]) […] When you specify a line like: DEVICE=…\display.sys con=(ega,437,1) […] DISPLAY.SYS is told that 437 should be assumed to be the hardware codepage of the display adapter. This is necessary to be known by DISPLAY.SYS because it will use different mechanisms to switch to the software prepared codepages and the hardware codepage, and there is no way to retrieve the info about the hardware codepage by itself. When the requested codepage is equal to the hardware codepage declared here, DISPLAY.SYS will use the method to switch to the hardware codepage, otherwise it will attempt to look up the font info in its internal buffer. […]
(NB. Based on fd-dev post .) "Changing code pages". XyWrite III, XyWrite III Plus and Signature. 2009-04-25 [2002-09-04]. Application Note #111. Archived from the original on 2016-06-12. Retrieved 2016-06-12. Elliott, John C. (2020-06-08). "Monochrome Display Adapter: Notes". Seasip.info. Archived from the original on 2023-09-20. Retrieved 2016-11-23. […] Reading from port 03BAh returns vertical sync in bit 7, and a card ID in bits 6-4: 000: Hercules, 001: Hercules Plus, 101: Hercules InColor, 111: Unknown clone […] One card […] which appears to be a Hercules clone of Eastern European origin, has a further capability - two ROM fonts, selectable in software. These options are controlled in a similar manner to other extended features in the Hercules: Bit 2 of port 3BFh must be set to enable the feature, and then Bit 4 of port 3B8h selects which font to use. The card ID in bits 6-4 of port 03BAh has all three bits set. […]
HP 95LX User's Guide (PDF) (2 ed.). Corvallis, OR, USA: Hewlett-Packard Company, Corvallis Division. June 1991 [March 1991]. pp. E-1–E-3, F-1–F-7. F0001-90003. Archived (PDF) from the original on 2016-11-28. Retrieved 2016-11-27. The HP 95LX character set is equivalent to code page 850, the IBM PC multilingual character set. (Note that your HP 95LX contains MS-DOS 3.22, which does not support code-page switching.) All the HP 95LX applications use this set except 1-2-3, which uses LICS, the Lotus International Character Set. Most LICS characters are included in code page 850; the few that are not will not display […] If your HP 95LX cannot display […] or if your printer cannot print a LICS character, the HP 95LX uses a fallback presentation for that character […] if you use the © symbol and your printer cannot print it, the HP 95LX might display (c) or c as the fallback presentation (depending on the capabilities of your printer).
Paul, Matthias R. (2002-12-04). "DISPLAY CON". FreeDOS development list freedos-dev. Archived from the original on 2017-09-09. Retrieved 2016-06-07. NEC Pinwriter. Ein Maßstab in der Profiklasse. (Printed 11-page color flyer about NEC Pinwriters P20/P30, P60/P70 and P90) (in German) (P-EAM-D-5/92 ed.), NEC Deutschland GmbH, May 1992 HP 200LX User's Guide (PDF) (1 ed.). Singapore: Hewlett-Packard Singapore (Private) Limited, Asia-Pacific Personal Computer Division. October 1996. pp. 21-6–21-7, 25-9–25-13, 26-14–26-18, C-1–C-4, D-1–D-10. 1216-90001. Archived (PDF) from the original on 2016-11-30. Retrieved 2016-11-29. Your palmtop supports code pages 850 and 437. The built-in applications (except 1-2-3) and System Manager-compliant applications […] use code page 850. […] Any DOS application you load and run will by default use code page 437. If an application requires code page 850, you can select it using the Setup utility.
Pinwriter P90 Bedienungshandbuch (Printed manual) (in German) (00 11/90 ed.). NEC Deutschland GmbH. November 1990. Programmierung des GQ-Modus für Epson-Seitendrucker (Printed manual) (in German) (2 ed.). Epson Deutschland GmbH. 1991. Pinwriter Familie - Pinwriter - Epromsockel - Zusätzliche Zeichensätze / Schriftarten (Printed reference manual for optional font and code page EPROMs for NEC Pinwriters, including custom variants) (in German) (00 3/93 ed.). NEC Deutschland GmbH. March 1993. bhd (2016-06-04) [1995]. "FreeBSD Handbook". The FreeBSD Documentation Project. 48895. Chernov, Andrey. 22.4.1. Localization - i18n/L10n Usage and Setup: Locale Configuration for Specific Languages: Russian Language (KOI8-R Encoding). Archived from the original on 2016-06-12. Retrieved 2016-06-12. bhd (2016-06-04) [1995]. "FreeBSD Handbook". The FreeBSD Documentation Project. 48895. Chernov, Andrey. 22.5.1.3 Localization - i18n/L10n Usage and Setup: Localizing FreeBSD to Specific Languages: Russian Language (KOI8-R Encoding): Printer Setup. Archived from the original on 2016-06-12. Retrieved 2016-06-12. Paul, Matthias R. (2001-04-09). NWDOS-TIPs — Tips & Tricks rund um Novell DOS 7, mit Blick auf undokumentierte Details, Bugs und Workarounds. Release 183 (in German) (3 ed.). MPDOSTIP.
Elliott, John C. (2006-10-14). "CPI file format". Seasip.info. Archived from the original on 2016-09-22. Retrieved 2016-09-22.
